Memory_retriever.ipynb

{
  "cells": [
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "dVUc1hWGRI-t"
      },
      "outputs": [],
      "source": []
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "Vht8wJ6fMUO2"
      },
      "outputs": [],
      "source": [
        "\n",
        "google_cloud_credentials = {\n",
        "  \"type\": \"service_account\",\n",
        "  \"project_id\": \"protean-cistern-407019\",\n",
        "  \"private_key_id\": \"cf35127f4cdc9cebf904f40ddac585414fe667e4\",\n",
        "  \"private_key\": \"-----BEGIN PRIVATE KEY-----\\nMIIEvgIBADANBgkqhkiG9w0BAQEFAASCBKgwggSkAgEAAoIBAQDCze5k6/wAWIDw\\nPv2M2RQ8tXWHJxBRkHz6+xYrGDCWMlEDK8M+RJF7D15nUWgqxZP0gzKO5xpeu0LM\\ns1BhYMlHCSxnjArre//rOokwVtI/B8y7RewEtckzOPak3D6N4g4/hX7iwj4NcTAX\\nvn8RwgXih2Xn7GnPj1KAA7olavNW66VOP06cUn+tvzMYPxqQlRBj6lilMx2UOWKw\\n4McxlgBzCEl3hHpPe+dd16JPnG4/TUusdLEg7do5ajTcgldK/DN/vP4DJ2LjyKfK\\nVOvkddZ6O3QuFMKw0FhcAh7xRIoVi4PMgmgI2jetPPte2/xaWN840wWLlq1Xht2K\\nKBLwqPz9AgMBAAECggEAB21oHT91IoM9WUuhibT+JUlq434NDHsUlNIteHOp6JQf\\nDB6nSVC+dkFyAUD0lorldwFolFPB6OI9bAYbNrdEvVlfcMxBX1ShP5QCBo6M4hvX\\nRb2m/8/5jaMqnkMKfIjlr6zUl2s+B++snH8mI6Rw6n/slgDo0OHfTKO3vXT8NbBh\\nwo6IPow+WjHwK9dIJbtZtQ4wXcRYVhargssVgMSOIFI2+nWN9jloyxlD52Mk817h\\n/9LuWyh2k5kz27PeDRm9uDozYuoqvyTZJJH70HxZeViYXt/BbVbYprtf0u6x7CSC\\nUzMqtEU8S/nmyYsJ8njuvpQPR/64P/NUW8lReV2/GQKBgQD4vQxgLjWSXIsky9DH\\nCq6TXdeLpL7cA7R68cv2jF2tcm0UjJI6308n3cAAK8h5ehTTYev7R9Pzsi9rY7Nb\\nnjibZpdzd7NB8mFlGUk9XMb4sRCS/+igZxzlh8c3YjK1P7w+/KBSEvSmjbCxVYiY\\nPG5VXP9mIk7/BSs+UCvy007v9QKBgQDIfc5EjsvSTU5ZKWxN1eo77PU53FgSSkM5\\nTSVUniUZkMV9Bo2zNIKZse3IIjwqZu5pNaquGpYFkIy6EA39fySvYi1KS8ZUSljP\\npa+qsPRBbNoTGlXM/M8V8u/uHCSe1ZycJ+pKrZkUhVauIw8OnsLKc/yeGXHdowEm\\nyICEWCLb6QKBgQC0pTeTsnudrDAxrQQXSh63T/P2qtMiHl6421px2EdAbFXkEqMw\\ntnVaUqXaefHCp3LwgbFMcUHxPntJ1sKvyxz5X2ymo0xAnLM3JUIXmoW2n4GWAN0F\\ngV1vJUj9ZAfHsl4YsIgOk02ymDy2zZPCXEEVJdEmb8xHUY2yV69TTzuI8QKBgQCm\\nZkvg+MdPtxu0wr2B6G4s6GErcaajH4GluFKoZU0TZ+WQ+/R1bLqMR3G7rjZTVvS+\\n22zabXRgD/XMo3CJGZJgAdLnG/bylHYbeT9QIufGuUpRmoTlylkxtVzPOgpmzZHa\\nxQIekvA/aN9lTfG8R6YSE6vxp9qqghR9lOGdde0QqQKBgBl04W3PWqZihXaFcUgp\\n4kFV1q5xzx/b1TMD+TwyOt409GkD5QXg+vYk4FZbXQT4Xs741ydOaNBtTt/R5TpL\\ncjVvOE9caziQClr6JIn1PyVNFI1pdPTfxFDNfRq8CFZEx4+FJEt0AqnsUAa4gdUX\\nE2JMNEPgs8/nuX7cdahe+eNK\\n-----END PRIVATE KEY-----\\n\",\n",
        "  \"client_email\": \"mml-research@protean-cistern-407019.iam.gserviceaccount.com\",\n",
        "  \"client_id\": \"109700191452046895861\",\n",
        "  \"auth_uri\": \"https://accounts.google.com/o/oauth2/auth\",\n",
        "  \"token_uri\": \"https://oauth2.googleapis.com/token\",\n",
        "  \"auth_provider_x509_cert_url\": \"https://www.googleapis.com/oauth2/v1/certs\",\n",
        "  \"client_x509_cert_url\": \"https://www.googleapis.com/robot/v1/metadata/x509/mml-research%40protean-cistern-407019.iam.gserviceaccount.com\",\n",
        "  \"universe_domain\": \"googleapis.com\"\n",
        "}\n",
        "ocr_service = GoogleVisionOCRService(google_cloud_credentials)\n",
        "tarsier = Tarsier(ocr_service)\n",
        "\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "R_JCh82b0z84",
        "outputId": "bc3adb95-7011-4c81-adf9-29e811691762"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Requirement already satisfied: playwright in /usr/local/lib/python3.10/dist-packages (1.40.0)\n",
            "Requirement already satisfied: greenlet==3.0.1 in /usr/local/lib/python3.10/dist-packages (from playwright) (3.0.1)\n",
            "Requirement already satisfied: pyee==11.0.1 in /usr/local/lib/python3.10/dist-packages (from playwright) (11.0.1)\n",
            "Requirement already satisfied: typing-extensions in /usr/local/lib/python3.10/dist-packages (from pyee==11.0.1->playwright) (4.5.0)\n"
          ]
        }
      ],
      "source": [
        "!pip install playwright"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "VLy8rFDR1EOX",
        "outputId": "0aec7d9e-4c53-497a-c633-876af743b281"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Downloading Chromium 120.0.6099.28 (playwright build v1091)\u001b[2m from https://playwright.azureedge.net/builds/chromium/1091/chromium-linux.zip\u001b[22m\n",
            "\u001b[1G153.1 Mb [] 0% 0.0s\u001b[0K\u001b[1G153.1 Mb [] 0% 14.3s\u001b[0K\u001b[1G153.1 Mb [] 0% 10.4s\u001b[0K\u001b[1G153.1 Mb [] 0% 8.7s\u001b[0K\u001b[1G153.1 Mb [] 0% 8.0s\u001b[0K\u001b[1G153.1 Mb [] 1% 7.2s\u001b[0K\u001b[1G153.1 Mb [] 1% 6.7s\u001b[0K\u001b[1G153.1 Mb [] 2% 6.5s\u001b[0K\u001b[1G153.1 Mb [] 2% 6.7s\u001b[0K\u001b[1G153.1 Mb [] 2% 6.9s\u001b[0K\u001b[1G153.1 Mb [] 2% 6.6s\u001b[0K\u001b[1G153.1 Mb [] 3% 6.7s\u001b[0K\u001b[1G153.1 Mb [] 3% 6.4s\u001b[0K\u001b[1G153.1 Mb [] 3% 6.6s\u001b[0K\u001b[1G153.1 Mb [] 3% 6.5s\u001b[0K\u001b[1G153.1 Mb [] 4% 6.4s\u001b[0K\u001b[1G153.1 Mb [] 4% 6.6s\u001b[0K\u001b[1G153.1 Mb [] 4% 6.4s\u001b[0K\u001b[1G153.1 Mb [] 5% 6.4s\u001b[0K\u001b[1G153.1 Mb [] 5% 6.2s\u001b[0K\u001b[1G153.1 Mb [] 5% 6.0s\u001b[0K\u001b[1G153.1 Mb [] 6% 6.1s\u001b[0K\u001b[1G153.1 Mb [] 6% 5.9s\u001b[0K\u001b[1G153.1 Mb [] 6% 5.8s\u001b[0K\u001b[1G153.1 Mb [] 6% 6.0s\u001b[0K\u001b[1G153.1 Mb [] 7% 5.9s\u001b[0K\u001b[1G153.1 Mb [] 7% 5.8s\u001b[0K\u001b[1G153.1 Mb [] 8% 5.7s\u001b[0K\u001b[1G153.1 Mb [] 8% 5.6s\u001b[0K\u001b[1G153.1 Mb [] 9% 5.4s\u001b[0K\u001b[1G153.1 Mb [] 9% 5.1s\u001b[0K\u001b[1G153.1 Mb [] 10% 5.0s\u001b[0K\u001b[1G153.1 Mb [] 10% 4.8s\u001b[0K\u001b[1G153.1 Mb [] 11% 4.7s\u001b[0K\u001b[1G153.1 Mb [] 11% 4.6s\u001b[0K\u001b[1G153.1 Mb [] 12% 4.7s\u001b[0K\u001b[1G153.1 Mb [] 12% 4.6s\u001b[0K\u001b[1G153.1 Mb [] 13% 4.4s\u001b[0K\u001b[1G153.1 Mb [] 14% 4.3s\u001b[0K\u001b[1G153.1 Mb [] 15% 4.3s\u001b[0K\u001b[1G153.1 Mb [] 15% 4.2s\u001b[0K\u001b[1G153.1 Mb [] 16% 4.1s\u001b[0K\u001b[1G153.1 Mb [] 17% 4.1s\u001b[0K\u001b[1G153.1 Mb [] 18% 4.1s\u001b[0K\u001b[1G153.1 Mb [] 18% 4.0s\u001b[0K\u001b[1G153.1 Mb [] 19% 3.9s\u001b[0K\u001b[1G153.1 Mb [] 19% 3.8s\u001b[0K\u001b[1G153.1 Mb [] 20% 3.8s\u001b[0K\u001b[1G153.1 Mb [] 21% 3.7s\u001b[0K\u001b[1G153.1 Mb [] 21% 3.6s\u001b[0K\u001b[1G153.1 Mb [] 22% 3.6s\u001b[0K\u001b[1G153.1 Mb [] 23% 3.5s\u001b[0K\u001b[1G153.1 Mb [] 23% 3.4s\u001b[0K\u001b[1G153.1 Mb [] 24% 3.4s\u001b[0K\u001b[1G153.1 Mb [] 24% 3.3s\u001b[0K\u001b[1G153.1 Mb [] 25% 3.3s\u001b[0K\u001b[1G153.1 Mb [] 26% 3.2s\u001b[0K\u001b[1G153.1 Mb [] 27% 3.2s\u001b[0K\u001b[1G153.1 Mb [] 28% 3.1s\u001b[0K\u001b[1G153.1 Mb [] 29% 3.2s\u001b[0K\u001b[1G153.1 Mb [] 29% 3.1s\u001b[0K\u001b[1G153.1 Mb [] 30% 3.1s\u001b[0K\u001b[1G153.1 Mb [] 30% 3.0s\u001b[0K\u001b[1G153.1 Mb [] 31% 2.9s\u001b[0K\u001b[1G153.1 Mb [] 32% 2.9s\u001b[0K\u001b[1G153.1 Mb [] 33% 2.8s\u001b[0K\u001b[1G153.1 Mb [] 34% 2.8s\u001b[0K\u001b[1G153.1 Mb [] 35% 2.7s\u001b[0K\u001b[1G153.1 Mb [] 36% 2.7s\u001b[0K\u001b[1G153.1 Mb [] 37% 2.6s\u001b[0K\u001b[1G153.1 Mb [] 38% 2.5s\u001b[0K\u001b[1G153.1 Mb [] 39% 2.4s\u001b[0K\u001b[1G153.1 Mb [] 41% 2.3s\u001b[0K\u001b[1G153.1 Mb [] 42% 2.2s\u001b[0K\u001b[1G153.1 Mb [] 43% 2.1s\u001b[0K\u001b[1G153.1 Mb [] 44% 2.0s\u001b[0K\u001b[1G153.1 Mb [] 45% 2.0s\u001b[0K\u001b[1G153.1 Mb [] 47% 1.9s\u001b[0K\u001b[1G153.1 Mb [] 48% 1.8s\u001b[0K\u001b[1G153.1 Mb [] 49% 1.7s\u001b[0K\u001b[1G153.1 Mb [] 50% 1.7s\u001b[0K\u001b[1G153.1 Mb [] 52% 1.6s\u001b[0K\u001b[1G153.1 Mb [] 53% 1.5s\u001b[0K\u001b[1G153.1 Mb [] 54% 1.5s\u001b[0K\u001b[1G153.1 Mb [] 56% 1.4s\u001b[0K\u001b[1G153.1 Mb [] 57% 1.3s\u001b[0K\u001b[1G153.1 Mb [] 58% 1.3s\u001b[0K\u001b[1G153.1 Mb [] 59% 1.2s\u001b[0K\u001b[1G153.1 Mb [] 60% 1.2s\u001b[0K\u001b[1G153.1 Mb [] 61% 1.2s\u001b[0K\u001b[1G153.1 Mb [] 62% 1.2s\u001b[0K\u001b[1G153.1 Mb [] 63% 1.1s\u001b[0K\u001b[1G153.1 Mb [] 64% 1.1s\u001b[0K\u001b[1G153.1 Mb [] 65% 1.0s\u001b[0K\u001b[1G153.1 Mb [] 66% 1.0s\u001b[0K\u001b[1G153.1 Mb [] 67% 0.9s\u001b[0K\u001b[1G153.1 Mb [] 69% 0.9s\u001b[0K\u001b[1G153.1 Mb [] 70% 0.8s\u001b[0K\u001b[1G153.1 Mb [] 71% 0.8s\u001b[0K\u001b[1G153.1 Mb [] 73% 0.7s\u001b[0K\u001b[1G153.1 Mb [] 74% 0.7s\u001b[0K\u001b[1G153.1 Mb [] 75% 0.7s\u001b[0K\u001b[1G153.1 Mb [] 76% 0.6s\u001b[0K\u001b[1G153.1 Mb [] 77% 0.6s\u001b[0K\u001b[1G153.1 Mb [] 79% 0.6s\u001b[0K\u001b[1G153.1 Mb [] 80% 0.5s\u001b[0K\u001b[1G153.1 Mb [] 81% 0.5s\u001b[0K\u001b[1G153.1 Mb [] 82% 0.4s\u001b[0K\u001b[1G153.1 Mb [] 84% 0.4s\u001b[0K\u001b[1G153.1 Mb [] 85% 0.4s\u001b[0K\u001b[1G153.1 Mb [] 86% 0.3s\u001b[0K\u001b[1G153.1 Mb [] 87% 0.3s\u001b[0K\u001b[1G153.1 Mb [] 88% 0.3s\u001b[0K\u001b[1G153.1 Mb [] 89% 0.3s\u001b[0K\u001b[1G153.1 Mb [] 90% 0.2s\u001b[0K\u001b[1G153.1 Mb [] 91% 0.2s\u001b[0K\u001b[1G153.1 Mb [] 92% 0.2s\u001b[0K\u001b[1G153.1 Mb [] 94% 0.1s\u001b[0K\u001b[1G153.1 Mb [] 95% 0.1s\u001b[0K\u001b[1G153.1 Mb [] 96% 0.1s\u001b[0K\u001b[1G153.1 Mb [] 97% 0.1s\u001b[0K\u001b[1G153.1 Mb [] 98% 0.0s\u001b[0K\u001b[1G153.1 Mb [] 99% 0.0s\u001b[0K\u001b[1G153.1 Mb [] 100% 0.0s\u001b[0K\n",
            "Chromium 120.0.6099.28 (playwright build v1091) downloaded to /root/.cache/ms-playwright/chromium-1091\n",
            "Downloading FFMPEG playwright build v1009\u001b[2m from https://playwright.azureedge.net/builds/ffmpeg/1009/ffmpeg-linux.zip\u001b[22m\n",
            "\u001b[1G2.6 Mb [] 0% 0.0s\u001b[0K\u001b[1G2.6 Mb [] 11% 0.1s\u001b[0K\u001b[1G2.6 Mb [] 31% 0.1s\u001b[0K\u001b[1G2.6 Mb [] 72% 0.0s\u001b[0K\u001b[1G2.6 Mb [] 100% 0.0s\u001b[0K\n",
            "FFMPEG playwright build v1009 downloaded to /root/.cache/ms-playwright/ffmpeg-1009\n",
            "Downloading Firefox 119.0 (playwright build v1429)\u001b[2m from https://playwright.azureedge.net/builds/firefox/1429/firefox-ubuntu-22.04.zip\u001b[22m\n",
            "\u001b[1G80.9 Mb [] 0% 0.0s\u001b[0K\u001b[1G80.9 Mb [] 0% 2.2s\u001b[0K\u001b[1G80.9 Mb [] 1% 1.8s\u001b[0K\u001b[1G80.9 Mb [] 3% 1.6s\u001b[0K\u001b[1G80.9 Mb [] 4% 1.5s\u001b[0K\u001b[1G80.9 Mb [] 5% 1.6s\u001b[0K\u001b[1G80.9 Mb [] 7% 1.5s\u001b[0K\u001b[1G80.9 Mb [] 8% 1.5s\u001b[0K\u001b[1G80.9 Mb [] 8% 1.6s\u001b[0K\u001b[1G80.9 Mb [] 9% 1.5s\u001b[0K\u001b[1G80.9 Mb [] 9% 2.8s\u001b[0K\u001b[1G80.9 Mb [] 10% 2.7s\u001b[0K\u001b[1G80.9 Mb [] 11% 2.5s\u001b[0K\u001b[1G80.9 Mb [] 12% 2.4s\u001b[0K\u001b[1G80.9 Mb [] 14% 2.2s\u001b[0K\u001b[1G80.9 Mb [] 15% 2.1s\u001b[0K\u001b[1G80.9 Mb [] 16% 2.1s\u001b[0K\u001b[1G80.9 Mb [] 18% 1.9s\u001b[0K\u001b[1G80.9 Mb [] 19% 4.0s\u001b[0K\u001b[1G80.9 Mb [] 20% 3.9s\u001b[0K\u001b[1G80.9 Mb [] 21% 3.8s\u001b[0K\u001b[1G80.9 Mb [] 22% 3.5s\u001b[0K\u001b[1G80.9 Mb [] 23% 3.5s\u001b[0K\u001b[1G80.9 Mb [] 24% 3.3s\u001b[0K\u001b[1G80.9 Mb [] 26% 3.1s\u001b[0K\u001b[1G80.9 Mb [] 27% 2.9s\u001b[0K\u001b[1G80.9 Mb [] 29% 2.7s\u001b[0K\u001b[1G80.9 Mb [] 31% 2.5s\u001b[0K\u001b[1G80.9 Mb [] 33% 2.4s\u001b[0K\u001b[1G80.9 Mb [] 35% 2.2s\u001b[0K\u001b[1G80.9 Mb [] 37% 2.0s\u001b[0K\u001b[1G80.9 Mb [] 38% 2.0s\u001b[0K\u001b[1G80.9 Mb [] 40% 1.8s\u001b[0K\u001b[1G80.9 Mb [] 42% 1.7s\u001b[0K\u001b[1G80.9 Mb [] 45% 1.6s\u001b[0K\u001b[1G80.9 Mb [] 46% 1.5s\u001b[0K\u001b[1G80.9 Mb [] 47% 1.4s\u001b[0K\u001b[1G80.9 Mb [] 49% 1.4s\u001b[0K\u001b[1G80.9 Mb [] 51% 1.3s\u001b[0K\u001b[1G80.9 Mb [] 53% 1.2s\u001b[0K\u001b[1G80.9 Mb [] 55% 1.1s\u001b[0K\u001b[1G80.9 Mb [] 58% 1.0s\u001b[0K\u001b[1G80.9 Mb [] 59% 1.0s\u001b[0K\u001b[1G80.9 Mb [] 62% 0.9s\u001b[0K\u001b[1G80.9 Mb [] 64% 0.8s\u001b[0K\u001b[1G80.9 Mb [] 66% 0.7s\u001b[0K\u001b[1G80.9 Mb [] 68% 0.7s\u001b[0K\u001b[1G80.9 Mb [] 70% 0.6s\u001b[0K\u001b[1G80.9 Mb [] 71% 0.6s\u001b[0K\u001b[1G80.9 Mb [] 72% 0.6s\u001b[0K\u001b[1G80.9 Mb [] 74% 0.5s\u001b[0K\u001b[1G80.9 Mb [] 77% 0.5s\u001b[0K\u001b[1G80.9 Mb [] 79% 0.4s\u001b[0K\u001b[1G80.9 Mb [] 81% 0.4s\u001b[0K\u001b[1G80.9 Mb [] 83% 0.3s\u001b[0K\u001b[1G80.9 Mb [] 86% 0.3s\u001b[0K\u001b[1G80.9 Mb [] 88% 0.2s\u001b[0K\u001b[1G80.9 Mb [] 90% 0.2s\u001b[0K\u001b[1G80.9 Mb [] 91% 0.2s\u001b[0K\u001b[1G80.9 Mb [] 91% 0.1s\u001b[0K\u001b[1G80.9 Mb [] 93% 0.1s\u001b[0K\u001b[1G80.9 Mb [] 94% 0.1s\u001b[0K\u001b[1G80.9 Mb [] 95% 0.1s\u001b[0K\u001b[1G80.9 Mb [] 97% 0.1s\u001b[0K\u001b[1G80.9 Mb [] 98% 0.0s\u001b[0K\u001b[1G80.9 Mb [] 99% 0.0s\u001b[0K\u001b[1G80.9 Mb [] 100% 0.0s\u001b[0K\n",
            "Firefox 119.0 (playwright build v1429) downloaded to /root/.cache/ms-playwright/firefox-1429\n",
            "Downloading Webkit 17.4 (playwright build v1944)\u001b[2m from https://playwright.azureedge.net/builds/webkit/1944/webkit-ubuntu-22.04.zip\u001b[22m\n",
            "\u001b[1G83.7 Mb [] 0% 0.0s\u001b[0K\u001b[1G83.7 Mb [] 0% 17.2s\u001b[0K\u001b[1G83.7 Mb [] 0% 5.3s\u001b[0K\u001b[1G83.7 Mb [] 1% 5.1s\u001b[0K\u001b[1G83.7 Mb [] 1% 4.9s\u001b[0K\u001b[1G83.7 Mb [] 1% 4.8s\u001b[0K\u001b[1G83.7 Mb [] 2% 4.0s\u001b[0K\u001b[1G83.7 Mb [] 3% 3.7s\u001b[0K\u001b[1G83.7 Mb [] 3% 3.8s\u001b[0K\u001b[1G83.7 Mb [] 4% 3.7s\u001b[0K\u001b[1G83.7 Mb [] 4% 3.8s\u001b[0K\u001b[1G83.7 Mb [] 5% 3.5s\u001b[0K\u001b[1G83.7 Mb [] 5% 3.7s\u001b[0K\u001b[1G83.7 Mb [] 6% 3.5s\u001b[0K\u001b[1G83.7 Mb [] 7% 3.5s\u001b[0K\u001b[1G83.7 Mb [] 8% 3.3s\u001b[0K\u001b[1G83.7 Mb [] 9% 3.2s\u001b[0K\u001b[1G83.7 Mb [] 10% 3.0s\u001b[0K\u001b[1G83.7 Mb [] 11% 2.9s\u001b[0K\u001b[1G83.7 Mb [] 12% 2.9s\u001b[0K\u001b[1G83.7 Mb [] 13% 2.9s\u001b[0K\u001b[1G83.7 Mb [] 14% 2.8s\u001b[0K\u001b[1G83.7 Mb [] 14% 2.9s\u001b[0K\u001b[1G83.7 Mb [] 15% 2.8s\u001b[0K\u001b[1G83.7 Mb [] 16% 2.7s\u001b[0K\u001b[1G83.7 Mb [] 17% 2.6s\u001b[0K\u001b[1G83.7 Mb [] 18% 2.6s\u001b[0K\u001b[1G83.7 Mb [] 19% 2.5s\u001b[0K\u001b[1G83.7 Mb [] 19% 2.4s\u001b[0K\u001b[1G83.7 Mb [] 20% 2.4s\u001b[0K\u001b[1G83.7 Mb [] 21% 2.3s\u001b[0K\u001b[1G83.7 Mb [] 22% 2.3s\u001b[0K\u001b[1G83.7 Mb [] 23% 2.3s\u001b[0K\u001b[1G83.7 Mb [] 24% 2.2s\u001b[0K\u001b[1G83.7 Mb [] 25% 2.2s\u001b[0K\u001b[1G83.7 Mb [] 26% 2.1s\u001b[0K\u001b[1G83.7 Mb [] 27% 2.1s\u001b[0K\u001b[1G83.7 Mb [] 28% 2.1s\u001b[0K\u001b[1G83.7 Mb [] 29% 2.0s\u001b[0K\u001b[1G83.7 Mb [] 30% 2.0s\u001b[0K\u001b[1G83.7 Mb [] 31% 2.0s\u001b[0K\u001b[1G83.7 Mb [] 32% 2.0s\u001b[0K\u001b[1G83.7 Mb [] 33% 1.9s\u001b[0K\u001b[1G83.7 Mb [] 33% 2.0s\u001b[0K\u001b[1G83.7 Mb [] 34% 1.9s\u001b[0K\u001b[1G83.7 Mb [] 35% 1.8s\u001b[0K\u001b[1G83.7 Mb [] 36% 1.8s\u001b[0K\u001b[1G83.7 Mb [] 37% 1.7s\u001b[0K\u001b[1G83.7 Mb [] 38% 1.7s\u001b[0K\u001b[1G83.7 Mb [] 39% 1.6s\u001b[0K\u001b[1G83.7 Mb [] 40% 1.6s\u001b[0K\u001b[1G83.7 Mb [] 41% 1.6s\u001b[0K\u001b[1G83.7 Mb [] 42% 1.5s\u001b[0K\u001b[1G83.7 Mb [] 43% 1.5s\u001b[0K\u001b[1G83.7 Mb [] 44% 1.4s\u001b[0K\u001b[1G83.7 Mb [] 45% 1.4s\u001b[0K\u001b[1G83.7 Mb [] 46% 1.4s\u001b[0K\u001b[1G83.7 Mb [] 47% 1.4s\u001b[0K\u001b[1G83.7 Mb [] 48% 1.3s\u001b[0K\u001b[1G83.7 Mb [] 49% 1.3s\u001b[0K\u001b[1G83.7 Mb [] 51% 1.2s\u001b[0K\u001b[1G83.7 Mb [] 52% 1.2s\u001b[0K\u001b[1G83.7 Mb [] 53% 1.2s\u001b[0K\u001b[1G83.7 Mb [] 54% 1.1s\u001b[0K\u001b[1G83.7 Mb [] 56% 1.1s\u001b[0K\u001b[1G83.7 Mb [] 56% 1.0s\u001b[0K\u001b[1G83.7 Mb [] 58% 1.0s\u001b[0K\u001b[1G83.7 Mb [] 59% 1.0s\u001b[0K\u001b[1G83.7 Mb [] 60% 0.9s\u001b[0K\u001b[1G83.7 Mb [] 62% 0.9s\u001b[0K\u001b[1G83.7 Mb [] 64% 0.8s\u001b[0K\u001b[1G83.7 Mb [] 65% 0.8s\u001b[0K\u001b[1G83.7 Mb [] 66% 0.8s\u001b[0K\u001b[1G83.7 Mb [] 67% 0.7s\u001b[0K\u001b[1G83.7 Mb [] 68% 0.7s\u001b[0K\u001b[1G83.7 Mb [] 69% 0.7s\u001b[0K\u001b[1G83.7 Mb [] 71% 0.6s\u001b[0K\u001b[1G83.7 Mb [] 72% 0.6s\u001b[0K\u001b[1G83.7 Mb [] 74% 0.6s\u001b[0K\u001b[1G83.7 Mb [] 75% 0.5s\u001b[0K\u001b[1G83.7 Mb [] 77% 0.5s\u001b[0K\u001b[1G83.7 Mb [] 78% 0.5s\u001b[0K\u001b[1G83.7 Mb [] 79% 0.4s\u001b[0K\u001b[1G83.7 Mb [] 80% 0.4s\u001b[0K\u001b[1G83.7 Mb [] 81% 0.4s\u001b[0K\u001b[1G83.7 Mb [] 83% 0.4s\u001b[0K\u001b[1G83.7 Mb [] 84% 0.3s\u001b[0K\u001b[1G83.7 Mb [] 85% 0.3s\u001b[0K\u001b[1G83.7 Mb [] 87% 0.2s\u001b[0K\u001b[1G83.7 Mb [] 89% 0.2s\u001b[0K\u001b[1G83.7 Mb [] 91% 0.2s\u001b[0K\u001b[1G83.7 Mb [] 92% 0.1s\u001b[0K\u001b[1G83.7 Mb [] 94% 0.1s\u001b[0K\u001b[1G83.7 Mb [] 95% 0.1s\u001b[0K\u001b[1G83.7 Mb [] 96% 0.1s\u001b[0K\u001b[1G83.7 Mb [] 98% 0.0s\u001b[0K\u001b[1G83.7 Mb [] 100% 0.0s\u001b[0K\n",
            "Webkit 17.4 (playwright build v1944) downloaded to /root/.cache/ms-playwright/webkit-1944\n"
          ]
        }
      ],
      "source": [
        "!playwright install"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "wiXpBZfJ0goP",
        "outputId": "c043127c-a738-4838-9cfc-61c67e5dae78"
      },
      "outputs": [],
      "source": [
        "async with async_playwright() as p:\n",
        "    browser = await p.chromium.launch(headless=True)\n",
        "    page = await browser.new_page()\n",
        "    await page.goto(\"http://shop.junglegym.ai//6s-wireless-headphones-over-ear-noise-canceling-hi-fi-bass-foldable-stereo-wireless-kid-headsets-earbuds-with-built-in-mic-micro-sd-tf-fm-for-iphone-samsung-ipad-pc-black-gold.html\")\n",
        "\n",
        "    page_text, tag_to_xpath = await tarsier.page_to_text(page)\n",
        "\n",
        "    # print(tag_to_xpath)  # Mapping of tags to x_paths\n",
        "    # print(page_text)  # My Text representation of the page"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "U5Uv-T8H1n6y",
        "outputId": "7b68936f-7eec-426e-a7c1-c6c947bdfc3e"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "{0: '//a[@id=\"idYYEhX8of\"]', 1: '//body[@id=\"html-body\"]/div[2][contains(concat(\" \", normalize-space(@class), \" \"), \" page-wrapper \")]/header[contains(concat(\" \", normalize-space(@class), \" \"), \" page-header \")]/div[1][contains(concat(\" \", normalize-space(@class), \" \"), \" panel \") and contains(concat(\" \", normalize-space(@class), \" \"), \" wrapper \")]/div[contains(concat(\" \", normalize-space(@class), \" \"), \" panel \") and contains(concat(\" \", normalize-space(@class), \" \"), \" header \")]/ul[contains(concat(\" \", normalize-space(@class), \" \"), \" header \") and contains(concat(\" \", normalize-space(@class), \" \"), \" links \")]/li[2][contains(concat(\" \", normalize-space(@class), \" \"), \" link \") and contains(concat(\" \", normalize-space(@class), \" \"), \" wishlist \")]/a', 2: '//body[@id=\"html-body\"]/div[2][contains(concat(\" \", normalize-space(@class), \" \"), \" page-wrapper \")]/header[contains(concat(\" \", normalize-space(@class), \" \"), \" page-header \")]/div[1][contains(concat(\" \", normalize-space(@class), \" \"), \" panel \") and contains(concat(\" \", normalize-space(@class), \" \"), \" wrapper \")]/div[contains(concat(\" \", normalize-space(@class), \" \"), \" panel \") and contains(concat(\" \", normalize-space(@class), \" \"), \" header \")]/ul[contains(concat(\" \", normalize-space(@class), \" \"), \" header \") and contains(concat(\" \", normalize-space(@class), \" \"), \" links \")]/li[3][contains(concat(\" \", normalize-space(@class), \" \"), \" link \") and contains(concat(\" \", normalize-space(@class), \" \"), \" authorization-link \")]/a', 3: '//a[@id=\"idFeAvHGE8\"]', 4: '//body[@id=\"html-body\"]/div[2][contains(concat(\" \", normalize-space(@class), \" \"), \" page-wrapper \")]/header[contains(concat(\" \", normalize-space(@class), \" \"), \" page-header \")]/div[1][contains(concat(\" \", normalize-space(@class), \" \"), \" panel \") and contains(concat(\" \", normalize-space(@class), \" \"), \" wrapper \")]/div[contains(concat(\" \", normalize-space(@class), \" \"), \" panel \") and contains(concat(\" \", normalize-space(@class), \" \"), \" header \")]/a[contains(concat(\" \", normalize-space(@class), \" \"), \" action \") and contains(concat(\" \", normalize-space(@class), \" \"), \" skip \") and contains(concat(\" \", normalize-space(@class), \" \"), \" contentarea \")]', 5: '//body[@id=\"html-body\"]/div[2][contains(concat(\" \", normalize-space(@class), \" \"), \" page-wrapper \")]/header[contains(concat(\" \", normalize-space(@class), \" \"), \" page-header \")]/div[2][contains(concat(\" \", normalize-space(@class), \" \"), \" header \") and contains(concat(\" \", normalize-space(@class), \" \"), \" content \")]/a[contains(concat(\" \", normalize-space(@class), \" \"), \" logo \")]', 6: '//body[@id=\"html-body\"]/div[2][contains(concat(\" \", normalize-space(@class), \" \"), \" page-wrapper \")]/header[contains(concat(\" \", normalize-space(@class), \" \"), \" page-header \")]/div[2][contains(concat(\" \", normalize-space(@class), \" \"), \" header \") and contains(concat(\" \", normalize-space(@class), \" \"), \" content \")]/div[1][contains(concat(\" \", normalize-space(@class), \" \"), \" minicart-wrapper \")]/a[contains(concat(\" \", normalize-space(@class), \" \"), \" action \") and contains(concat(\" \", normalize-space(@class), \" \"), \" showcart \")]', 7: '//form[@id=\"search_mini_form\"]/div[1][contains(concat(\" \", normalize-space(@class), \" \"), \" field \") and contains(concat(\" \", normalize-space(@class), \" \"), \" search \")]/label[contains(concat(\" \", normalize-space(@class), \" \"), \" label \")]', 8: '//input[@id=\"search\"]', 9: '//form[@id=\"search_mini_form\"]/div[1][contains(concat(\" \", normalize-space(@class), \" \"), \" field \") and contains(concat(\" \", normalize-space(@class), \" \"), \" search \")]/div[contains(concat(\" \", normalize-space(@class), \" \"), \" control \")]/div[2][contains(concat(\" \", normalize-space(@class), \" \"), \" nested \")]/a[contains(concat(\" \", normalize-space(@class), \" \"), \" action \") and contains(concat(\" \", normalize-space(@class), \" \"), \" advanced \")]', 10: '//form[@id=\"search_mini_form\"]/div[2][contains(concat(\" \", normalize-space(@class), \" \"), \" actions \")]/button[contains(concat(\" \", normalize-space(@class), \" \"), \" action \") and contains(concat(\" \", normalize-space(@class), \" \"), \" search \")]', 11: '//div[2][@id=\"store.menu\"]/nav[contains(concat(\" \", normalize-space(@class), \" \"), \" navigation \")]/ul/li[1][contains(concat(\" \", normalize-space(@class), \" \"), \" level0 \") and contains(concat(\" \", normalize-space(@class), \" \"), \" nav-1 \") and contains(concat(\" \", normalize-space(@class), \" \"), \" category-item \") and contains(concat(\" \", normalize-space(@class), \" \"), \" first \") and contains(concat(\" \", normalize-space(@class), \" \"), \" level-top \") and contains(concat(\" \", normalize-space(@class), \" \"), \" parent \")]/a[contains(concat(\" \", normalize-space(@class), \" \"), \" level-top \")]', 12: '//div[2][@id=\"store.menu\"]/nav[contains(concat(\" \", normalize-space(@class), \" \"), \" navigation \")]/ul/li[2][contains(concat(\" \", normalize-space(@class), \" \"), \" level0 \") and contains(concat(\" \", normalize-space(@class), \" \"), \" nav-2 \") and contains(concat(\" \", normalize-space(@class), \" \"), \" category-item \") and contains(concat(\" \", normalize-space(@class), \" \"), \" level-top \") and contains(concat(\" \", normalize-space(@class), \" \"), \" parent \")]/a[contains(concat(\" \", normalize-space(@class), \" \"), \" level-top \")]', 13: '//div[2][@id=\"store.menu\"]/nav[contains(concat(\" \", normalize-space(@class), \" \"), \" navigation \")]/ul/li[3][contains(concat(\" \", normalize-space(@class), \" \"), \" level0 \") and contains(concat(\" \", normalize-space(@class), \" \"), \" nav-3 \") and contains(concat(\" \", normalize-space(@class), \" \"), \" category-item \") and contains(concat(\" \", normalize-space(@class), \" \"), \" level-top \") and contains(concat(\" \", normalize-space(@class), \" \"), \" parent \")]/a[contains(concat(\" \", normalize-space(@class), \" \"), \" level-top \")]', 14: '//div[2][@id=\"store.menu\"]/nav[contains(concat(\" \", normalize-space(@class), \" \"), \" navigation \")]/ul/li[4][contains(concat(\" \", normalize-space(@class), \" \"), \" level0 \") and contains(concat(\" \", normalize-space(@class), \" \"), \" nav-4 \") and contains(concat(\" \", normalize-space(@class), \" \"), \" category-item \") and contains(concat(\" \", normalize-space(@class), \" \"), \" level-top \") and contains(concat(\" \", normalize-space(@class), \" \"), \" parent \")]/a[contains(concat(\" \", normalize-space(@class), \" \"), \" level-top \")]', 15: '//div[2][@id=\"store.menu\"]/nav[contains(concat(\" \", normalize-space(@class), \" \"), \" navigation \")]/ul/li[5][contains(concat(\" \", normalize-space(@class), \" \"), \" level0 \") and contains(concat(\" \", normalize-space(@class), \" \"), \" nav-5 \") and contains(concat(\" \", normalize-space(@class), \" \"), \" category-item \") and contains(concat(\" \", normalize-space(@class), \" \"), \" level-top \") and contains(concat(\" \", normalize-space(@class), \" \"), \" parent \")]/a[contains(concat(\" \", normalize-space(@class), \" \"), \" level-top \")]', 16: '//div[2][@id=\"store.menu\"]/nav[contains(concat(\" \", normalize-space(@class), \" \"), \" navigation \")]/ul/li[6][contains(concat(\" \", normalize-space(@class), \" \"), \" level0 \") and contains(concat(\" \", normalize-space(@class), \" \"), \" nav-6 \") and contains(concat(\" \", normalize-space(@class), \" \"), \" category-item \") and contains(concat(\" \", normalize-space(@class), \" \"), \" level-top \") and contains(concat(\" \", normalize-space(@class), \" \"), \" parent \")]/a[contains(concat(\" \", normalize-space(@class), \" \"), \" level-top \")]', 17: '//div[2][@id=\"store.menu\"]/nav[contains(concat(\" \", normalize-space(@class), \" \"), \" navigation \")]/ul/li[7][contains(concat(\" \", normalize-space(@class), \" \"), \" level0 \") and contains(concat(\" \", normalize-space(@class), \" \"), \" nav-7 \") and contains(concat(\" \", normalize-space(@class), \" \"), \" category-item \") and contains(concat(\" \", normalize-space(@class), \" \"), \" level-top \") and contains(concat(\" \", normalize-space(@class), \" \"), \" parent \")]/a[contains(concat(\" \", normalize-space(@class), \" \"), \" level-top \")]', 18: '//div[2][@id=\"store.menu\"]/nav[contains(concat(\" \", normalize-space(@class), \" \"), \" navigation \")]/ul/li[8][contains(concat(\" \", normalize-space(@class), \" \"), \" level0 \") and contains(concat(\" \", normalize-space(@class), \" \"), \" nav-8 \") and contains(concat(\" \", normalize-space(@class), \" \"), \" category-item \") and contains(concat(\" \", normalize-space(@class), \" \"), \" level-top \") and contains(concat(\" \", normalize-space(@class), \" \"), \" parent \")]/a[contains(concat(\" \", normalize-space(@class), \" \"), \" level-top \")]', 19: '//div[2][@id=\"store.menu\"]/nav[contains(concat(\" \", normalize-space(@class), \" \"), \" navigation \")]/ul/li[9][contains(concat(\" \", normalize-space(@class), \" \"), \" level0 \") and contains(concat(\" \", normalize-space(@class), \" \"), \" nav-9 \") and contains(concat(\" \", normalize-space(@class), \" \"), \" category-item \") and contains(concat(\" \", normalize-space(@class), \" \"), \" level-top \") and contains(concat(\" \", normalize-space(@class), \" \"), \" parent \")]/a[contains(concat(\" \", normalize-space(@class), \" \"), \" level-top \")]', 20: '//div[2][@id=\"store.menu\"]/nav[contains(concat(\" \", normalize-space(@class), \" \"), \" navigation \")]/ul/li[10][contains(concat(\" \", normalize-space(@class), \" \"), \" level0 \") and contains(concat(\" \", normalize-space(@class), \" \"), \" nav-10 \") and contains(concat(\" \", normalize-space(@class), \" \"), \" category-item \") and contains(concat(\" \", normalize-space(@class), \" \"), \" level-top \") and contains(concat(\" \", normalize-space(@class), \" \"), \" parent \")]/a[contains(concat(\" \", normalize-space(@class), \" \"), \" level-top \")]', 21: '//div[2][@id=\"store.menu\"]/nav[contains(concat(\" \", normalize-space(@class), \" \"), \" navigation \")]/ul/li[11][contains(concat(\" \", normalize-space(@class), \" \"), \" level0 \") and contains(concat(\" \", normalize-space(@class), \" \"), \" nav-11 \") and contains(concat(\" \", normalize-space(@class), \" \"), \" category-item \") and contains(concat(\" \", normalize-space(@class), \" \"), \" level-top \") and contains(concat(\" \", normalize-space(@class), \" \"), \" parent \")]/a[contains(concat(\" \", normalize-space(@class), \" \"), \" level-top \")]', 22: '//div[2][@id=\"store.menu\"]/nav[contains(concat(\" \", normalize-space(@class), \" \"), \" navigation \")]/ul/li[12][contains(concat(\" \", normalize-space(@class), \" \"), \" level0 \") and contains(concat(\" \", normalize-space(@class), \" \"), \" nav-12 \") and contains(concat(\" \", normalize-space(@class), \" \"), \" category-item \") and contains(concat(\" \", normalize-space(@class), \" \"), \" last \") and contains(concat(\" \", normalize-space(@class), \" \"), \" level-top \") and contains(concat(\" \", normalize-space(@class), \" \"), \" parent \")]/a[contains(concat(\" \", normalize-space(@class), \" \"), \" level-top \")]', 23: '//main[@id=\"maincontent\"]/div[3][contains(concat(\" \", normalize-space(@class), \" \"), \" columns \")]/div[contains(concat(\" \", normalize-space(@class), \" \"), \" column \") and contains(concat(\" \", normalize-space(@class), \" \"), \" main \")]/div[1][contains(concat(\" \", normalize-space(@class), \" \"), \" product-info-main \")]/div[1][contains(concat(\" \", normalize-space(@class), \" \"), \" product-info-price \")]/div[2][contains(concat(\" \", normalize-space(@class), \" \"), \" product-reviews-summary \")]/div[2][contains(concat(\" \", normalize-space(@class), \" \"), \" reviews-actions \")]/a[1][contains(concat(\" \", normalize-space(@class), \" \"), \" action \") and contains(concat(\" \", normalize-space(@class), \" \"), \" view \")]', 24: '//main[@id=\"maincontent\"]/div[3][contains(concat(\" \", normalize-space(@class), \" \"), \" columns \")]/div[contains(concat(\" \", normalize-space(@class), \" \"), \" column \") and contains(concat(\" \", normalize-space(@class), \" \"), \" main \")]/div[1][contains(concat(\" \", normalize-space(@class), \" \"), \" product-info-main \")]/div[1][contains(concat(\" \", normalize-space(@class), \" \"), \" product-info-price \")]/div[2][contains(concat(\" \", normalize-space(@class), \" \"), \" product-reviews-summary \")]/div[2][contains(concat(\" \", normalize-space(@class), \" \"), \" reviews-actions \")]/a[2][contains(concat(\" \", normalize-space(@class), \" \"), \" action \") and contains(concat(\" \", normalize-space(@class), \" \"), \" add \")]', 25: '//div[1][@id=\"product-options-wrapper\"]/div[contains(concat(\" \", normalize-space(@class), \" \"), \" fieldset \")]/div[contains(concat(\" \", normalize-space(@class), \" \"), \" field \") and contains(concat(\" \", normalize-space(@class), \" \"), \" required \")]/label[contains(concat(\" \", normalize-space(@class), \" \"), \" label \")]', 26: '//input[@id=\"options_51745_2\"]', 27: '//div[@id=\"options-51745-list\"]/div[1][contains(concat(\" \", normalize-space(@class), \" \"), \" field \") and contains(concat(\" \", normalize-space(@class), \" \"), \" choice \") and contains(concat(\" \", normalize-space(@class), \" \"), \" admin__field \") and contains(concat(\" \", normalize-space(@class), \" \"), \" admin__field-option \")]/label[contains(concat(\" \", normalize-space(@class), \" \"), \" label \") and contains(concat(\" \", normalize-space(@class), \" \"), \" admin__field-label \")]', 28: '//input[@id=\"options_51745_3\"]', 29: '//div[@id=\"options-51745-list\"]/div[2][contains(concat(\" \", normalize-space(@class), \" \"), \" field \") and contains(concat(\" \", normalize-space(@class), \" \"), \" choice \") and contains(concat(\" \", normalize-space(@class), \" \"), \" admin__field \") and contains(concat(\" \", normalize-space(@class), \" \"), \" admin__field-option \")]/label[contains(concat(\" \", normalize-space(@class), \" \"), \" label \") and contains(concat(\" \", normalize-space(@class), \" \"), \" admin__field-label \")]', 30: '//input[@id=\"options_51745_4\"]', 31: '//div[@id=\"options-51745-list\"]/div[3][contains(concat(\" \", normalize-space(@class), \" \"), \" field \") and contains(concat(\" \", normalize-space(@class), \" \"), \" choice \") and contains(concat(\" \", normalize-space(@class), \" \"), \" admin__field \") and contains(concat(\" \", normalize-space(@class), \" \"), \" admin__field-option \")]/label[contains(concat(\" \", normalize-space(@class), \" \"), \" label \") and contains(concat(\" \", normalize-space(@class), \" \"), \" admin__field-label \")]', 32: '//input[@id=\"options_51745_5\"]', 33: '//div[@id=\"options-51745-list\"]/div[4][contains(concat(\" \", normalize-space(@class), \" \"), \" field \") and contains(concat(\" \", normalize-space(@class), \" \"), \" choice \") and contains(concat(\" \", normalize-space(@class), \" \"), \" admin__field \") and contains(concat(\" \", normalize-space(@class), \" \"), \" admin__field-option \")]/label[contains(concat(\" \", normalize-space(@class), \" \"), \" label \") and contains(concat(\" \", normalize-space(@class), \" \"), \" admin__field-label \")]', 34: '//input[@id=\"options_51745_6\"]', 35: '//div[@id=\"options-51745-list\"]/div[5][contains(concat(\" \", normalize-space(@class), \" \"), \" field \") and contains(concat(\" \", normalize-space(@class), \" \"), \" choice \") and contains(concat(\" \", normalize-space(@class), \" \"), \" admin__field \") and contains(concat(\" \", normalize-space(@class), \" \"), \" admin__field-option \")]/label[contains(concat(\" \", normalize-space(@class), \" \"), \" label \") and contains(concat(\" \", normalize-space(@class), \" \"), \" admin__field-label \")]', 36: '//form[@id=\"product_addtocart_form\"]/div[2][contains(concat(\" \", normalize-space(@class), \" \"), \" product-options-bottom \")]/div[contains(concat(\" \", normalize-space(@class), \" \"), \" box-tocart \")]/div[contains(concat(\" \", normalize-space(@class), \" \"), \" fieldset \")]/div[1][contains(concat(\" \", normalize-space(@class), \" \"), \" field \") and contains(concat(\" \", normalize-space(@class), \" \"), \" qty \")]/label[contains(concat(\" \", normalize-space(@class), \" \"), \" label \")]', 37: '//input[@id=\"qty\"]', 38: '//main[@id=\"maincontent\"]/div[3][contains(concat(\" \", normalize-space(@class), \" \"), \" columns \")]/div[contains(concat(\" \", normalize-space(@class), \" \"), \" column \") and contains(concat(\" \", normalize-space(@class), \" \"), \" main \")]/div[1][contains(concat(\" \", normalize-space(@class), \" \"), \" product-info-main \")]/div[3][contains(concat(\" \", normalize-space(@class), \" \"), \" product-social-links \")]/div[contains(concat(\" \", normalize-space(@class), \" \"), \" product-addto-links \")]/a[1][contains(concat(\" \", normalize-space(@class), \" \"), \" action \") and contains(concat(\" \", normalize-space(@class), \" \"), \" towishlist \")]', 39: '//main[@id=\"maincontent\"]/div[3][contains(concat(\" \", normalize-space(@class), \" \"), \" columns \")]/div[contains(concat(\" \", normalize-space(@class), \" \"), \" column \") and contains(concat(\" \", normalize-space(@class), \" \"), \" main \")]/div[1][contains(concat(\" \", normalize-space(@class), \" \"), \" product-info-main \")]/div[3][contains(concat(\" \", normalize-space(@class), \" \"), \" product-social-links \")]/div[contains(concat(\" \", normalize-space(@class), \" \"), \" product-addto-links \")]/a[2][contains(concat(\" \", normalize-space(@class), \" \"), \" action \") and contains(concat(\" \", normalize-space(@class), \" \"), \" tocompare \")]', 40: '//main[@id=\"maincontent\"]/div[3][contains(concat(\" \", normalize-space(@class), \" \"), \" columns \")]/div[contains(concat(\" \", normalize-space(@class), \" \"), \" column \") and contains(concat(\" \", normalize-space(@class), \" \"), \" main \")]/div[2][contains(concat(\" \", normalize-space(@class), \" \"), \" product \") and contains(concat(\" \", normalize-space(@class), \" \"), \" media \")]/div[1][contains(concat(\" \", normalize-space(@class), \" \"), \" action-skip-wrapper \")]/a[contains(concat(\" \", normalize-space(@class), \" \"), \" action \") and contains(concat(\" \", normalize-space(@class), \" \"), \" skip \") and contains(concat(\" \", normalize-space(@class), \" \"), \" gallery-next-area \")]', 41: '//main[@id=\"maincontent\"]/div[3][contains(concat(\" \", normalize-space(@class), \" \"), \" columns \")]/div[contains(concat(\" \", normalize-space(@class), \" \"), \" column \") and contains(concat(\" \", normalize-space(@class), \" \"), \" main \")]/div[2][contains(concat(\" \", normalize-space(@class), \" \"), \" product \") and contains(concat(\" \", normalize-space(@class), \" \"), \" media \")]/div[3][contains(concat(\" \", normalize-space(@class), \" \"), \" action-skip-wrapper \")]/a[contains(concat(\" \", normalize-space(@class), \" \"), \" action \") and contains(concat(\" \", normalize-space(@class), \" \"), \" skip \") and contains(concat(\" \", normalize-space(@class), \" \"), \" gallery-prev-area \")]', 42: '//a[@id=\"tab-label-description-title\"]', 43: '//a[@id=\"tab-label-reviews-title\"]', 44: '//form[@id=\"newsletter-validate-detail\"]/div[1][contains(concat(\" \", normalize-space(@class), \" \"), \" field \") and contains(concat(\" \", normalize-space(@class), \" \"), \" newsletter \")]/div[contains(concat(\" \", normalize-space(@class), \" \"), \" control \")]/label', 45: '//form[@id=\"newsletter-validate-detail\"]/div[2][contains(concat(\" \", normalize-space(@class), \" \"), \" actions \")]/button[contains(concat(\" \", normalize-space(@class), \" \"), \" action \") and contains(concat(\" \", normalize-space(@class), \" \"), \" subscribe \") and contains(concat(\" \", normalize-space(@class), \" \"), \" primary \")]', 46: '//body[@id=\"html-body\"]/div[2][contains(concat(\" \", normalize-space(@class), \" \"), \" page-wrapper \")]/footer[contains(concat(\" \", normalize-space(@class), \" \"), \" page-footer \")]/div[contains(concat(\" \", normalize-space(@class), \" \"), \" footer \") and contains(concat(\" \", normalize-space(@class), \" \"), \" content \")]/ul[contains(concat(\" \", normalize-space(@class), \" \"), \" footer \") and contains(concat(\" \", normalize-space(@class), \" \"), \" links \")]/li[1][contains(concat(\" \", normalize-space(@class), \" \"), \" nav \") and contains(concat(\" \", normalize-space(@class), \" \"), \" item \")]/a', 47: '//body[@id=\"html-body\"]/div[2][contains(concat(\" \", normalize-space(@class), \" \"), \" page-wrapper \")]/footer[contains(concat(\" \", normalize-space(@class), \" \"), \" page-footer \")]/div[contains(concat(\" \", normalize-space(@class), \" \"), \" footer \") and contains(concat(\" \", normalize-space(@class), \" \"), \" content \")]/ul[contains(concat(\" \", normalize-space(@class), \" \"), \" footer \") and contains(concat(\" \", normalize-space(@class), \" \"), \" links \")]/li[2][contains(concat(\" \", normalize-space(@class), \" \"), \" nav \") and contains(concat(\" \", normalize-space(@class), \" \"), \" item \")]/a', 48: '//body[@id=\"html-body\"]/div[2][contains(concat(\" \", normalize-space(@class), \" \"), \" page-wrapper \")]/footer[contains(concat(\" \", normalize-space(@class), \" \"), \" page-footer \")]/div[contains(concat(\" \", normalize-space(@class), \" \"), \" footer \") and contains(concat(\" \", normalize-space(@class), \" \"), \" content \")]/ul[contains(concat(\" \", normalize-space(@class), \" \"), \" footer \") and contains(concat(\" \", normalize-space(@class), \" \"), \" links \")]/li[3][contains(concat(\" \", normalize-space(@class), \" \"), \" nav \") and contains(concat(\" \", normalize-space(@class), \" \"), \" item \")]/a', 49: '//body[@id=\"html-body\"]/div[2][contains(concat(\" \", normalize-space(@class), \" \"), \" page-wrapper \")]/footer[contains(concat(\" \", normalize-space(@class), \" \"), \" page-footer \")]/div[contains(concat(\" \", normalize-space(@class), \" \"), \" footer \") and contains(concat(\" \", normalize-space(@class), \" \"), \" content \")]/ul[contains(concat(\" \", normalize-space(@class), \" \"), \" footer \") and contains(concat(\" \", normalize-space(@class), \" \"), \" links \")]/li[4][contains(concat(\" \", normalize-space(@class), \" \"), \" nav \") and contains(concat(\" \", normalize-space(@class), \" \"), \" item \")]/a', 50: '//body[@id=\"html-body\"]/div[2][contains(concat(\" \", normalize-space(@class), \" \"), \" page-wrapper \")]/footer[contains(concat(\" \", normalize-space(@class), \" \"), \" page-footer \")]/div[contains(concat(\" \", normalize-space(@class), \" \"), \" footer \") and contains(concat(\" \", normalize-space(@class), \" \"), \" content \")]/ul[contains(concat(\" \", normalize-space(@class), \" \"), \" footer \") and contains(concat(\" \", normalize-space(@class), \" \"), \" links \")]/li[5][contains(concat(\" \", normalize-space(@class), \" \"), \" nav \") and contains(concat(\" \", normalize-space(@class), \" \"), \" item \")]/a', 51: '//body[@id=\"html-body\"]/div[2][contains(concat(\" \", normalize-space(@class), \" \"), \" page-wrapper \")]/footer[contains(concat(\" \", normalize-space(@class), \" \"), \" page-footer \")]/div[contains(concat(\" \", normalize-space(@class), \" \"), \" footer \") and contains(concat(\" \", normalize-space(@class), \" \"), \" content \")]/small[2][contains(concat(\" \", normalize-space(@class), \" \"), \" bugs \")]/a'}\n"
          ]
        }
      ],
      "source": [
        "print(tag_to_xpath)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "Q30fO7kH-Isn"
      },
      "outputs": [],
      "source": [
        "data_object = {\n",
        "    'prompt': \"\",\n",
        "    'img': \"image_path\",\n",
        "    'observation': \"\",\n",
        "    'action': \"\"\n",
        "\n",
        "}"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "4fPNyYo8zp9O",
        "outputId": "3c048b26-478b-4ad2-9e86-686c9f0e0dbd"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Requirement already satisfied: timm in /usr/local/lib/python3.10/dist-packages (0.9.12)\n",
            "Requirement already satisfied: torch>=1.7 in /usr/local/lib/python3.10/dist-packages (from timm) (2.1.0+cu118)\n",
            "Requirement already satisfied: torchvision in /usr/local/lib/python3.10/dist-packages (from timm) (0.16.0+cu118)\n",
            "Requirement already satisfied: pyyaml in /usr/local/lib/python3.10/dist-packages (from timm) (6.0.1)\n",
            "Requirement already satisfied: huggingface-hub in /usr/local/lib/python3.10/dist-packages (from timm) (0.19.4)\n",
            "Requirement already satisfied: safetensors in /usr/local/lib/python3.10/dist-packages (from timm) (0.4.1)\n",
            "Requirement already satisfied: filelock in /usr/local/lib/python3.10/dist-packages (from torch>=1.7->timm) (3.13.1)\n",
            "Requirement already satisfied: typing-extensions in /usr/local/lib/python3.10/dist-packages (from torch>=1.7->timm) (4.5.0)\n",
            "Requirement already satisfied: sympy in /usr/local/lib/python3.10/dist-packages (from torch>=1.7->timm) (1.12)\n",
            "Requirement already satisfied: networkx in /usr/local/lib/python3.10/dist-packages (from torch>=1.7->timm) (3.2.1)\n",
            "Requirement already satisfied: jinja2 in /usr/local/lib/python3.10/dist-packages (from torch>=1.7->timm) (3.1.2)\n",
            "Requirement already satisfied: fsspec in /usr/local/lib/python3.10/dist-packages (from torch>=1.7->timm) (2023.6.0)\n",
            "Requirement already satisfied: triton==2.1.0 in /usr/local/lib/python3.10/dist-packages (from torch>=1.7->timm) (2.1.0)\n",
            "Requirement already satisfied: requests in /usr/local/lib/python3.10/dist-packages (from huggingface-hub->timm) (2.31.0)\n",
            "Requirement already satisfied: tqdm>=4.42.1 in /usr/local/lib/python3.10/dist-packages (from huggingface-hub->timm) (4.66.1)\n",
            "Requirement already satisfied: packaging>=20.9 in /usr/local/lib/python3.10/dist-packages (from huggingface-hub->timm) (23.2)\n",
            "Requirement already satisfied: numpy in /usr/local/lib/python3.10/dist-packages (from torchvision->timm) (1.23.5)\n",
            "Requirement already satisfied: pillow!=8.3.*,>=5.3.0 in /usr/local/lib/python3.10/dist-packages (from torchvision->timm) (9.4.0)\n",
            "Requirement already satisfied: MarkupSafe>=2.0 in /usr/local/lib/python3.10/dist-packages (from jinja2->torch>=1.7->timm) (2.1.3)\n",
            "Requirement already satisfied: charset-normalizer<4,>=2 in /usr/local/lib/python3.10/dist-packages (from requests->huggingface-hub->timm) (3.3.2)\n",
            "Requirement already satisfied: idna<4,>=2.5 in /usr/local/lib/python3.10/dist-packages (from requests->huggingface-hub->timm) (3.6)\n",
            "Requirement already satisfied: urllib3<3,>=1.21.1 in /usr/local/lib/python3.10/dist-packages (from requests->huggingface-hub->timm) (2.0.7)\n",
            "Requirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python3.10/dist-packages (from requests->huggingface-hub->timm) (2023.11.17)\n",
            "Requirement already satisfied: mpmath>=0.19 in /usr/local/lib/python3.10/dist-packages (from sympy->torch>=1.7->timm) (1.3.0)\n",
            "Requirement already satisfied: transformers in /usr/local/lib/python3.10/dist-packages (4.35.2)\n",
            "Requirement already satisfied: filelock in /usr/local/lib/python3.10/dist-packages (from transformers) (3.13.1)\n",
            "Requirement already satisfied: huggingface-hub<1.0,>=0.16.4 in /usr/local/lib/python3.10/dist-packages (from transformers) (0.19.4)\n",
            "Requirement already satisfied: numpy>=1.17 in /usr/local/lib/python3.10/dist-packages (from transformers) (1.23.5)\n",
            "Requirement already satisfied: packaging>=20.0 in /usr/local/lib/python3.10/dist-packages (from transformers) (23.2)\n",
            "Requirement already satisfied: pyyaml>=5.1 in /usr/local/lib/python3.10/dist-packages (from transformers) (6.0.1)\n",
            "Requirement already satisfied: regex!=2019.12.17 in /usr/local/lib/python3.10/dist-packages (from transformers) (2023.6.3)\n",
            "Requirement already satisfied: requests in /usr/local/lib/python3.10/dist-packages (from transformers) (2.31.0)\n",
            "Requirement already satisfied: tokenizers<0.19,>=0.14 in /usr/local/lib/python3.10/dist-packages (from transformers) (0.15.0)\n",
            "Requirement already satisfied: safetensors>=0.3.1 in /usr/local/lib/python3.10/dist-packages (from transformers) (0.4.1)\n",
            "Requirement already satisfied: tqdm>=4.27 in /usr/local/lib/python3.10/dist-packages (from transformers) (4.66.1)\n",
            "Requirement already satisfied: fsspec>=2023.5.0 in /usr/local/lib/python3.10/dist-packages (from huggingface-hub<1.0,>=0.16.4->transformers) (2023.6.0)\n",
            "Requirement already satisfied: typing-extensions>=3.7.4.3 in /usr/local/lib/python3.10/dist-packages (from huggingface-hub<1.0,>=0.16.4->transformers) (4.5.0)\n",
            "Requirement already satisfied: charset-normalizer<4,>=2 in /usr/local/lib/python3.10/dist-packages (from requests->transformers) (3.3.2)\n",
            "Requirement already satisfied: idna<4,>=2.5 in /usr/local/lib/python3.10/dist-packages (from requests->transformers) (3.6)\n",
            "Requirement already satisfied: urllib3<3,>=1.21.1 in /usr/local/lib/python3.10/dist-packages (from requests->transformers) (2.0.7)\n",
            "Requirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python3.10/dist-packages (from requests->transformers) (2023.11.17)\n"
          ]
        }
      ],
      "source": [
        "!pip install timm\n",
        "!pip install transformers"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 363
        },
        "id": "WdN4SEmDcrbR",
        "outputId": "f2c3b35e-ab41-4262-c07d-aaf8c784d230"
      },
      "outputs": [
        {
          "ename": "KeyboardInterrupt",
          "evalue": "ignored",
          "output_type": "error",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mKeyboardInterrupt\u001b[0m                         Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-3-7244c023c75f>\u001b[0m in \u001b[0;36m<cell line: 6>\u001b[0;34m()\u001b[0m\n\u001b[1;32m      4\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0malbumentations\u001b[0m \u001b[0;32mas\u001b[0m \u001b[0mA\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      5\u001b[0m \u001b[0;32mfrom\u001b[0m \u001b[0mtorch\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0mnn\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 6\u001b[0;31m \u001b[0;32mimport\u001b[0m \u001b[0mtimm\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
            "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/timm/__init__.py\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m      1\u001b[0m \u001b[0;32mfrom\u001b[0m \u001b[0;34m.\u001b[0m\u001b[0mversion\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0m__version__\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 2\u001b[0;31m \u001b[0;32mfrom\u001b[0m \u001b[0;34m.\u001b[0m\u001b[0mlayers\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0mis_scriptable\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mis_exportable\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mset_scriptable\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mset_exportable\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      3\u001b[0m \u001b[0;32mfrom\u001b[0m \u001b[0;34m.\u001b[0m\u001b[0mmodels\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0mcreate_model\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mlist_models\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mlist_pretrained\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mis_model\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mlist_modules\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmodel_entrypoint\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;31m \u001b[0m\u001b[0;31m\\\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      4\u001b[0m     \u001b[0mis_model_pretrained\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mget_pretrained_cfg\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mget_pretrained_cfg_value\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/timm/layers/__init__.py\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m      5\u001b[0m \u001b[0;32mfrom\u001b[0m \u001b[0;34m.\u001b[0m\u001b[0mattention_pool2d\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0mAttentionPool2d\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mRotAttentionPool2d\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mRotaryEmbedding\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      6\u001b[0m \u001b[0;32mfrom\u001b[0m \u001b[0;34m.\u001b[0m\u001b[0mblur_pool\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0mBlurPool2d\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 7\u001b[0;31m \u001b[0;32mfrom\u001b[0m \u001b[0;34m.\u001b[0m\u001b[0mclassifier\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0mClassifierHead\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mcreate_classifier\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mNormMlpClassifierHead\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      8\u001b[0m \u001b[0;32mfrom\u001b[0m \u001b[0;34m.\u001b[0m\u001b[0mcond_conv2d\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0mCondConv2d\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mget_condconv_initializer\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      9\u001b[0m \u001b[0;32mfrom\u001b[0m \u001b[0;34m.\u001b[0m\u001b[0mconfig\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0mis_exportable\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mis_scriptable\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mis_no_jit\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0muse_fused_attn\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;31m \u001b[0m\u001b[0;31m\\\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/timm/layers/classifier.py\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m     13\u001b[0m \u001b[0;32mfrom\u001b[0m \u001b[0;34m.\u001b[0m\u001b[0madaptive_avgmax_pool\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0mSelectAdaptivePool2d\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     14\u001b[0m \u001b[0;32mfrom\u001b[0m \u001b[0;34m.\u001b[0m\u001b[0mcreate_act\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0mget_act_layer\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 15\u001b[0;31m \u001b[0;32mfrom\u001b[0m \u001b[0;34m.\u001b[0m\u001b[0mcreate_norm\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0mget_norm_layer\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     16\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     17\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/timm/layers/create_norm.py\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m     12\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     13\u001b[0m \u001b[0;32mfrom\u001b[0m \u001b[0;34m.\u001b[0m\u001b[0mnorm\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0mGroupNorm\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mGroupNorm1\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mLayerNorm\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mLayerNorm2d\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mRmsNorm\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 14\u001b[0;31m \u001b[0;32mfrom\u001b[0m \u001b[0mtorchvision\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mops\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmisc\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0mFrozenBatchNorm2d\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     15\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     16\u001b[0m _NORM_MAP = dict(\n",
            "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/torchvision/__init__.py\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m      4\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      5\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0mtorch\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 6\u001b[0;31m \u001b[0;32mfrom\u001b[0m \u001b[0mtorchvision\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0m_meta_registrations\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mdatasets\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mio\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mmodels\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mops\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtransforms\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mutils\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      7\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      8\u001b[0m \u001b[0;32mfrom\u001b[0m \u001b[0;34m.\u001b[0m\u001b[0mextension\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0m_HAS_OPS\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/torchvision/models/__init__.py\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m      1\u001b[0m \u001b[0;32mfrom\u001b[0m \u001b[0;34m.\u001b[0m\u001b[0malexnet\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 2\u001b[0;31m \u001b[0;32mfrom\u001b[0m \u001b[0;34m.\u001b[0m\u001b[0mconvnext\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      3\u001b[0m \u001b[0;32mfrom\u001b[0m \u001b[0;34m.\u001b[0m\u001b[0mdensenet\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      4\u001b[0m \u001b[0;32mfrom\u001b[0m \u001b[0;34m.\u001b[0m\u001b[0mefficientnet\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      5\u001b[0m \u001b[0;32mfrom\u001b[0m \u001b[0;34m.\u001b[0m\u001b[0mgooglenet\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0;34m*\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/torchvision/models/convnext.py\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m      6\u001b[0m \u001b[0;32mfrom\u001b[0m \u001b[0mtorch\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mnn\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0mfunctional\u001b[0m \u001b[0;32mas\u001b[0m \u001b[0mF\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      7\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 8\u001b[0;31m \u001b[0;32mfrom\u001b[0m \u001b[0;34m.\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mops\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmisc\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0mConv2dNormActivation\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mPermute\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      9\u001b[0m \u001b[0;32mfrom\u001b[0m \u001b[0;34m.\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mops\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mstochastic_depth\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0mStochasticDepth\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     10\u001b[0m \u001b[0;32mfrom\u001b[0m \u001b[0;34m.\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtransforms\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_presets\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0mImageClassification\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/torchvision/ops/__init__.py\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0;32mfrom\u001b[0m \u001b[0;34m.\u001b[0m\u001b[0m_register_onnx_ops\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0m_register_custom_op\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      2\u001b[0m from .boxes import (\n\u001b[1;32m      3\u001b[0m     \u001b[0mbatched_nms\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      4\u001b[0m     \u001b[0mbox_area\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      5\u001b[0m     \u001b[0mbox_convert\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/torchvision/ops/_register_onnx_ops.py\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m      3\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      4\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0mtorch\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 5\u001b[0;31m \u001b[0;32mfrom\u001b[0m \u001b[0mtorch\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0monnx\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0msymbolic_opset11\u001b[0m \u001b[0;32mas\u001b[0m \u001b[0mopset11\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      6\u001b[0m \u001b[0;32mfrom\u001b[0m \u001b[0mtorch\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0monnx\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0msymbolic_helper\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0mparse_args\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      7\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/torch/onnx/__init__.py\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m     44\u001b[0m )\n\u001b[1;32m     45\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 46\u001b[0;31m from ._internal.exporter import (  # usort:skip. needs to be last to avoid circular import\n\u001b[0m\u001b[1;32m     47\u001b[0m     \u001b[0mExportOptions\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     48\u001b[0m     \u001b[0mExportOutput\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/torch/onnx/_internal/exporter.py\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m     40\u001b[0m \u001b[0;32mfrom\u001b[0m \u001b[0mtorch\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0monnx\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0m_internal\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mdiagnostics\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0minfra\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     41\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 42\u001b[0;31m from torch.onnx._internal.fx import (\n\u001b[0m\u001b[1;32m     43\u001b[0m     \u001b[0mdecomposition_table\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     44\u001b[0m     \u001b[0mpatcher\u001b[0m \u001b[0;32mas\u001b[0m \u001b[0mpatcher\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/torch/onnx/_internal/fx/__init__.py\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0;32mfrom\u001b[0m \u001b[0;34m.\u001b[0m\u001b[0mpatcher\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0mONNXTorchPatcher\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      2\u001b[0m \u001b[0;32mfrom\u001b[0m \u001b[0;34m.\u001b[0m\u001b[0mserialization\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0msave_model_with_external_data\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      3\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      4\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      5\u001b[0m __all__ = [\n",
            "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/torch/onnx/_internal/fx/patcher.py\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m      8\u001b[0m \u001b[0;32mtry\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      9\u001b[0m     \u001b[0;31m# safetensors is not an exporter requirement, but needed for some huggingface models\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 10\u001b[0;31m     \u001b[0;32mimport\u001b[0m \u001b[0msafetensors\u001b[0m  \u001b[0;31m# type: ignore[import]  # noqa: F401\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m     11\u001b[0m     \u001b[0;32mimport\u001b[0m \u001b[0mtransformers\u001b[0m  \u001b[0;31m# type: ignore[import]\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     12\u001b[0m     \u001b[0;32mfrom\u001b[0m \u001b[0msafetensors\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0mtorch\u001b[0m \u001b[0;32mas\u001b[0m \u001b[0msafetensors_torch\u001b[0m  \u001b[0;31m# noqa: F401\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/safetensors/__init__.py\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m      1\u001b[0m \u001b[0;31m# Re-export this\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 2\u001b[0;31m from ._safetensors_rust import (  # noqa: F401\n\u001b[0m\u001b[1;32m      3\u001b[0m     \u001b[0mSafetensorError\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      4\u001b[0m     \u001b[0m__version__\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      5\u001b[0m     \u001b[0mdeserialize\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;31mKeyboardInterrupt\u001b[0m: "
          ]
        }
      ],
      "source": [
        "import os\n",
        "import cv2\n",
        "import torch\n",
        "import albumentations as A\n",
        "from torch import nn\n",
        "import timm\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "jj3zm8vGfkuz"
      },
      "source": [
        "#Memory Retrieval - Vector Similarity\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "jStmb-lKjtLw",
        "outputId": "cfb1f1f4-c7fa-4872-81a7-85146dec98cd"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Collecting timm\n",
            "  Downloading timm-0.9.12-py3-none-any.whl (2.2 MB)\n",
            "\u001b[2K     \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m2.2/2.2 MB\u001b[0m \u001b[31m13.8 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hRequirement already satisfied: torch>=1.7 in /usr/local/lib/python3.10/dist-packages (from timm) (2.1.0+cu118)\n",
            "Requirement already satisfied: torchvision in /usr/local/lib/python3.10/dist-packages (from timm) (0.16.0+cu118)\n",
            "Requirement already satisfied: pyyaml in /usr/local/lib/python3.10/dist-packages (from timm) (6.0.1)\n",
            "Requirement already satisfied: huggingface-hub in /usr/local/lib/python3.10/dist-packages (from timm) (0.19.4)\n",
            "Requirement already satisfied: safetensors in /usr/local/lib/python3.10/dist-packages (from timm) (0.4.1)\n",
            "Requirement already satisfied: filelock in /usr/local/lib/python3.10/dist-packages (from torch>=1.7->timm) (3.13.1)\n",
            "Requirement already satisfied: typing-extensions in /usr/local/lib/python3.10/dist-packages (from torch>=1.7->timm) (4.5.0)\n",
            "Requirement already satisfied: sympy in /usr/local/lib/python3.10/dist-packages (from torch>=1.7->timm) (1.12)\n",
            "Requirement already satisfied: networkx in /usr/local/lib/python3.10/dist-packages (from torch>=1.7->timm) (3.2.1)\n",
            "Requirement already satisfied: jinja2 in /usr/local/lib/python3.10/dist-packages (from torch>=1.7->timm) (3.1.2)\n",
            "Requirement already satisfied: fsspec in /usr/local/lib/python3.10/dist-packages (from torch>=1.7->timm) (2023.6.0)\n",
            "Requirement already satisfied: triton==2.1.0 in /usr/local/lib/python3.10/dist-packages (from torch>=1.7->timm) (2.1.0)\n",
            "Requirement already satisfied: requests in /usr/local/lib/python3.10/dist-packages (from huggingface-hub->timm) (2.31.0)\n",
            "Requirement already satisfied: tqdm>=4.42.1 in /usr/local/lib/python3.10/dist-packages (from huggingface-hub->timm) (4.66.1)\n",
            "Requirement already satisfied: packaging>=20.9 in /usr/local/lib/python3.10/dist-packages (from huggingface-hub->timm) (23.2)\n",
            "Requirement already satisfied: numpy in /usr/local/lib/python3.10/dist-packages (from torchvision->timm) (1.23.5)\n",
            "Requirement already satisfied: pillow!=8.3.*,>=5.3.0 in /usr/local/lib/python3.10/dist-packages (from torchvision->timm) (9.4.0)\n",
            "Requirement already satisfied: MarkupSafe>=2.0 in /usr/local/lib/python3.10/dist-packages (from jinja2->torch>=1.7->timm) (2.1.3)\n",
            "Requirement already satisfied: charset-normalizer<4,>=2 in /usr/local/lib/python3.10/dist-packages (from requests->huggingface-hub->timm) (3.3.2)\n",
            "Requirement already satisfied: idna<4,>=2.5 in /usr/local/lib/python3.10/dist-packages (from requests->huggingface-hub->timm) (3.6)\n",
            "Requirement already satisfied: urllib3<3,>=1.21.1 in /usr/local/lib/python3.10/dist-packages (from requests->huggingface-hub->timm) (2.0.7)\n",
            "Requirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python3.10/dist-packages (from requests->huggingface-hub->timm) (2023.11.17)\n",
            "Requirement already satisfied: mpmath>=0.19 in /usr/local/lib/python3.10/dist-packages (from sympy->torch>=1.7->timm) (1.3.0)\n",
            "Installing collected packages: timm\n",
            "Successfully installed timm-0.9.12\n",
            "Requirement already satisfied: transformers in /usr/local/lib/python3.10/dist-packages (4.35.2)\n",
            "Requirement already satisfied: filelock in /usr/local/lib/python3.10/dist-packages (from transformers) (3.13.1)\n",
            "Requirement already satisfied: huggingface-hub<1.0,>=0.16.4 in /usr/local/lib/python3.10/dist-packages (from transformers) (0.19.4)\n",
            "Requirement already satisfied: numpy>=1.17 in /usr/local/lib/python3.10/dist-packages (from transformers) (1.23.5)\n",
            "Requirement already satisfied: packaging>=20.0 in /usr/local/lib/python3.10/dist-packages (from transformers) (23.2)\n",
            "Requirement already satisfied: pyyaml>=5.1 in /usr/local/lib/python3.10/dist-packages (from transformers) (6.0.1)\n",
            "Requirement already satisfied: regex!=2019.12.17 in /usr/local/lib/python3.10/dist-packages (from transformers) (2023.6.3)\n",
            "Requirement already satisfied: requests in /usr/local/lib/python3.10/dist-packages (from transformers) (2.31.0)\n",
            "Requirement already satisfied: tokenizers<0.19,>=0.14 in /usr/local/lib/python3.10/dist-packages (from transformers) (0.15.0)\n",
            "Requirement already satisfied: safetensors>=0.3.1 in /usr/local/lib/python3.10/dist-packages (from transformers) (0.4.1)\n",
            "Requirement already satisfied: tqdm>=4.27 in /usr/local/lib/python3.10/dist-packages (from transformers) (4.66.1)\n",
            "Requirement already satisfied: fsspec>=2023.5.0 in /usr/local/lib/python3.10/dist-packages (from huggingface-hub<1.0,>=0.16.4->transformers) (2023.6.0)\n",
            "Requirement already satisfied: typing-extensions>=3.7.4.3 in /usr/local/lib/python3.10/dist-packages (from huggingface-hub<1.0,>=0.16.4->transformers) (4.5.0)\n",
            "Requirement already satisfied: charset-normalizer<4,>=2 in /usr/local/lib/python3.10/dist-packages (from requests->transformers) (3.3.2)\n",
            "Requirement already satisfied: idna<4,>=2.5 in /usr/local/lib/python3.10/dist-packages (from requests->transformers) (3.6)\n",
            "Requirement already satisfied: urllib3<3,>=1.21.1 in /usr/local/lib/python3.10/dist-packages (from requests->transformers) (2.0.7)\n",
            "Requirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python3.10/dist-packages (from requests->transformers) (2023.11.17)\n"
          ]
        }
      ],
      "source": [
        "!pip install timm\n",
        "!pip install transformers"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "abAYv2BlftA4"
      },
      "outputs": [],
      "source": [
        "import os\n",
        "import cv2\n",
        "import gc\n",
        "import numpy as np\n",
        "import pandas as pd\n",
        "import itertools\n",
        "from tqdm import tqdm\n",
        "import albumentations as A\n",
        "\n",
        "import torch\n",
        "from torch import nn\n",
        "import torch.nn.functional as F\n",
        "import timm\n",
        "from transformers import DistilBertModel, DistilBertConfig, DistilBertTokenizer"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "D0SoDAE64qPj"
      },
      "outputs": [],
      "source": [
        "#insert path to json file \n",
        "data = pd.read_json(\"\").T"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 424
        },
        "id": "Sg4ADMgF55zI",
        "outputId": "3424e61f-0d19-4c43-be5b-608259c38869"
      },
      "outputs": [
        {
          "data": {
            "text/html": [
              "\n",
              "  <div id=\"df-41e7b346-88e1-4234-8f34-678600f117b8\" class=\"colab-df-container\">\n",
              "    <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>image</th>\n",
              "      <th>prompt</th>\n",
              "      <th>actions</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>render_571.html</th>\n",
              "      <td>render_571.png</td>\n",
              "      <td>I recently moved, my address is 231 Willow Wa...</td>\n",
              "      <td>click [1256]</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>render_434.html</th>\n",
              "      <td>render_434.png</td>\n",
              "      <td>Add the product with the lowest per unit pric...</td>\n",
              "      <td>click [1323]</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>render_160.html</th>\n",
              "      <td>render_160.png</td>\n",
              "      <td>I have a lot of Nintendo Switch game cards no...</td>\n",
              "      <td>click [887]</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>render_530.html</th>\n",
              "      <td>render_530.png</td>\n",
              "      <td>Draft a refund message via their \"contact us\"...</td>\n",
              "      <td>click [1256]</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>render_588.html</th>\n",
              "      <td>render_588.png</td>\n",
              "      <td>Rate my recent purchase of Foundation For Mat...</td>\n",
              "      <td>click [1265]</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>render_227.html</th>\n",
              "      <td>render_227.png</td>\n",
              "      <td>What is the price range for products from EYZ...</td>\n",
              "      <td>goto http://homepage.com/password.html</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>render_335.html</th>\n",
              "      <td>render_335.png</td>\n",
              "      <td>Tell me when I last ordered my body butter?\\n</td>\n",
              "      <td>goto http://onestopmarket.com/my-account</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>render_270.html</th>\n",
              "      <td>render_270.png</td>\n",
              "      <td>Show me products under $30 in \"men shoes\" cat...</td>\n",
              "      <td>click [850]</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>render_159.html</th>\n",
              "      <td>render_159.png</td>\n",
              "      <td>I have a lot of Nintendo Switch game cards no...</td>\n",
              "      <td>click [887]</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>render_509.html</th>\n",
              "      <td>render_509.png</td>\n",
              "      <td>Buy the best rating product from \"Men's shoe\"...</td>\n",
              "      <td>click [895]</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>187 rows × 3 columns</p>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-41e7b346-88e1-4234-8f34-678600f117b8')\"\n",
              "            title=\"Convert this dataframe to an interactive table.\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "    <script>\n",
              "      const buttonEl =\n",
              "        document.querySelector('#df-41e7b346-88e1-4234-8f34-678600f117b8 button.colab-df-convert');\n",
              "      buttonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "      async function convertToInteractive(key) {\n",
              "        const element = document.querySelector('#df-41e7b346-88e1-4234-8f34-678600f117b8');\n",
              "        const dataTable =\n",
              "          await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                    [key], {});\n",
              "        if (!dataTable) return;\n",
              "\n",
              "        const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "          + ' to learn more about interactive tables.';\n",
              "        element.innerHTML = '';\n",
              "        dataTable['output_type'] = 'display_data';\n",
              "        await google.colab.output.renderOutput(dataTable, element);\n",
              "        const docLink = document.createElement('div');\n",
              "        docLink.innerHTML = docLinkHtml;\n",
              "        element.appendChild(docLink);\n",
              "      }\n",
              "    </script>\n",
              "  </div>\n",
              "\n",
              "\n",
              "<div id=\"df-a612bbde-6173-4042-baac-04f8c697d00b\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-a612bbde-6173-4042-baac-04f8c697d00b')\"\n",
              "            title=\"Suggest charts\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "      --bg-color: #E8F0FE;\n",
              "      --fill-color: #1967D2;\n",
              "      --hover-bg-color: #E2EBFA;\n",
              "      --hover-fill-color: #174EA6;\n",
              "      --disabled-fill-color: #AAA;\n",
              "      --disabled-bg-color: #DDD;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "      --bg-color: #3B4455;\n",
              "      --fill-color: #D2E3FC;\n",
              "      --hover-bg-color: #434B5C;\n",
              "      --hover-fill-color: #FFFFFF;\n",
              "      --disabled-bg-color: #3B4455;\n",
              "      --disabled-fill-color: #666;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart {\n",
              "    background-color: var(--bg-color);\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: var(--fill-color);\n",
              "    height: 32px;\n",
              "    padding: 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: var(--hover-bg-color);\n",
              "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: var(--button-hover-fill-color);\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart-complete:disabled,\n",
              "  .colab-df-quickchart-complete:disabled:hover {\n",
              "    background-color: var(--disabled-bg-color);\n",
              "    fill: var(--disabled-fill-color);\n",
              "    box-shadow: none;\n",
              "  }\n",
              "\n",
              "  .colab-df-spinner {\n",
              "    border: 2px solid var(--fill-color);\n",
              "    border-color: transparent;\n",
              "    border-bottom-color: var(--fill-color);\n",
              "    animation:\n",
              "      spin 1s steps(1) infinite;\n",
              "  }\n",
              "\n",
              "  @keyframes spin {\n",
              "    0% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "      border-left-color: var(--fill-color);\n",
              "    }\n",
              "    20% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    30% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    40% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    60% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    80% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "    90% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "  }\n",
              "</style>\n",
              "\n",
              "  <script>\n",
              "    async function quickchart(key) {\n",
              "      const quickchartButtonEl =\n",
              "        document.querySelector('#' + key + ' button');\n",
              "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
              "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
              "      try {\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      } catch (error) {\n",
              "        console.error('Error during call to suggestCharts:', error);\n",
              "      }\n",
              "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
              "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-a612bbde-6173-4042-baac-04f8c697d00b button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "    </div>\n",
              "  </div>\n"
            ],
            "text/plain": [
              "                          image  \\\n",
              "render_571.html  render_571.png   \n",
              "render_434.html  render_434.png   \n",
              "render_160.html  render_160.png   \n",
              "render_530.html  render_530.png   \n",
              "render_588.html  render_588.png   \n",
              "...                         ...   \n",
              "render_227.html  render_227.png   \n",
              "render_335.html  render_335.png   \n",
              "render_270.html  render_270.png   \n",
              "render_159.html  render_159.png   \n",
              "render_509.html  render_509.png   \n",
              "\n",
              "                                                            prompt  \\\n",
              "render_571.html   I recently moved, my address is 231 Willow Wa...   \n",
              "render_434.html   Add the product with the lowest per unit pric...   \n",
              "render_160.html   I have a lot of Nintendo Switch game cards no...   \n",
              "render_530.html   Draft a refund message via their \"contact us\"...   \n",
              "render_588.html   Rate my recent purchase of Foundation For Mat...   \n",
              "...                                                            ...   \n",
              "render_227.html   What is the price range for products from EYZ...   \n",
              "render_335.html      Tell me when I last ordered my body butter?\\n   \n",
              "render_270.html   Show me products under $30 in \"men shoes\" cat...   \n",
              "render_159.html   I have a lot of Nintendo Switch game cards no...   \n",
              "render_509.html   Buy the best rating product from \"Men's shoe\"...   \n",
              "\n",
              "                                                  actions  \n",
              "render_571.html                              click [1256]  \n",
              "render_434.html                              click [1323]  \n",
              "render_160.html                               click [887]  \n",
              "render_530.html                              click [1256]  \n",
              "render_588.html                              click [1265]  \n",
              "...                                                   ...  \n",
              "render_227.html    goto http://homepage.com/password.html  \n",
              "render_335.html  goto http://onestopmarket.com/my-account  \n",
              "render_270.html                               click [850]  \n",
              "render_159.html                               click [887]  \n",
              "render_509.html                               click [895]  \n",
              "\n",
              "[187 rows x 3 columns]"
            ]
          },
          "execution_count": 50,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "data"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "oWi0l4s-Jyuv"
      },
      "outputs": [],
      "source": [
        "data['caption'] = data['prompt'] + \": \" + data['actions']"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "wE0phRYpKJch"
      },
      "outputs": [],
      "source": [
        "#change the path to add the path to images\n",
        "data['img_path'] = \"Images/\"+data['image']"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 304
        },
        "id": "KOu-1DkMM6rT",
        "outputId": "6cef4302-93ab-4e6b-ad0a-a6fe0ecaeaee"
      },
      "outputs": [
        {
          "ename": "AttributeError",
          "evalue": "ignored",
          "output_type": "error",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mAttributeError\u001b[0m                            Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-9-a7b27b9f874c>\u001b[0m in \u001b[0;36m<cell line: 1>\u001b[0;34m()\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mremoved_data\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mpd\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mcopy\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mdata\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m",
            "\u001b[0;32m/usr/local/lib/python3.10/dist-packages/pandas/__init__.py\u001b[0m in \u001b[0;36m__getattr__\u001b[0;34m(name)\u001b[0m\n\u001b[1;32m    262\u001b[0m         \u001b[0;32mreturn\u001b[0m \u001b[0m_SparseArray\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    263\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m--> 264\u001b[0;31m     \u001b[0;32mraise\u001b[0m \u001b[0mAttributeError\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34mf\"module 'pandas' has no attribute '{name}'\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m    265\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m    266\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;31mAttributeError\u001b[0m: module 'pandas' has no attribute 'copy'"
          ]
        }
      ],
      "source": [
        "removed_data = pd.copy(data)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "fDc5R31NMTG8"
      },
      "outputs": [],
      "source": [
        "for i in data.index:\n",
        "  if not os.path.isfile(data['img_path'][i]):\n",
        "    data = data.drop(i)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "Ua3UjsWmKU8T"
      },
      "outputs": [],
      "source": [
        "for i in data.index:\n",
        "  print(os.path.isfile(data['img_path'][i]))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "lk_CG-orOqx9"
      },
      "outputs": [],
      "source": []
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "cCeEsG5MNyRT"
      },
      "outputs": [],
      "source": [
        "data = data.reset_index()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 788
        },
        "id": "ghNMm1T9OLxy",
        "outputId": "863df4c1-da71-46c0-e564-e9292fb1b358"
      },
      "outputs": [
        {
          "data": {
            "text/html": [
              "\n",
              "  <div id=\"df-d287b007-b4c6-4c67-bce8-c3167b2cd214\" class=\"colab-df-container\">\n",
              "    <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>index</th>\n",
              "      <th>image</th>\n",
              "      <th>prompt</th>\n",
              "      <th>actions</th>\n",
              "      <th>caption</th>\n",
              "      <th>img_path</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>render_571.html</td>\n",
              "      <td>render_571.png</td>\n",
              "      <td>I recently moved, my address is 231 Willow Wa...</td>\n",
              "      <td>click [1256]</td>\n",
              "      <td>I recently moved, my address is 231 Willow Wa...</td>\n",
              "      <td>/content/drive/MyDrive/MML Research/Images/ren...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>render_434.html</td>\n",
              "      <td>render_434.png</td>\n",
              "      <td>Add the product with the lowest per unit pric...</td>\n",
              "      <td>click [1323]</td>\n",
              "      <td>Add the product with the lowest per unit pric...</td>\n",
              "      <td>/content/drive/MyDrive/MML Research/Images/ren...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>render_160.html</td>\n",
              "      <td>render_160.png</td>\n",
              "      <td>I have a lot of Nintendo Switch game cards no...</td>\n",
              "      <td>click [887]</td>\n",
              "      <td>I have a lot of Nintendo Switch game cards no...</td>\n",
              "      <td>/content/drive/MyDrive/MML Research/Images/ren...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>render_530.html</td>\n",
              "      <td>render_530.png</td>\n",
              "      <td>Draft a refund message via their \"contact us\"...</td>\n",
              "      <td>click [1256]</td>\n",
              "      <td>Draft a refund message via their \"contact us\"...</td>\n",
              "      <td>/content/drive/MyDrive/MML Research/Images/ren...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>render_588.html</td>\n",
              "      <td>render_588.png</td>\n",
              "      <td>Rate my recent purchase of Foundation For Mat...</td>\n",
              "      <td>click [1265]</td>\n",
              "      <td>Rate my recent purchase of Foundation For Mat...</td>\n",
              "      <td>/content/drive/MyDrive/MML Research/Images/ren...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>179</th>\n",
              "      <td>render_227.html</td>\n",
              "      <td>render_227.png</td>\n",
              "      <td>What is the price range for products from EYZ...</td>\n",
              "      <td>goto http://homepage.com/password.html</td>\n",
              "      <td>What is the price range for products from EYZ...</td>\n",
              "      <td>/content/drive/MyDrive/MML Research/Images/ren...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>180</th>\n",
              "      <td>render_335.html</td>\n",
              "      <td>render_335.png</td>\n",
              "      <td>Tell me when I last ordered my body butter?\\n</td>\n",
              "      <td>goto http://onestopmarket.com/my-account</td>\n",
              "      <td>Tell me when I last ordered my body butter?\\n...</td>\n",
              "      <td>/content/drive/MyDrive/MML Research/Images/ren...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>181</th>\n",
              "      <td>render_270.html</td>\n",
              "      <td>render_270.png</td>\n",
              "      <td>Show me products under $30 in \"men shoes\" cat...</td>\n",
              "      <td>click [850]</td>\n",
              "      <td>Show me products under $30 in \"men shoes\" cat...</td>\n",
              "      <td>/content/drive/MyDrive/MML Research/Images/ren...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>182</th>\n",
              "      <td>render_159.html</td>\n",
              "      <td>render_159.png</td>\n",
              "      <td>I have a lot of Nintendo Switch game cards no...</td>\n",
              "      <td>click [887]</td>\n",
              "      <td>I have a lot of Nintendo Switch game cards no...</td>\n",
              "      <td>/content/drive/MyDrive/MML Research/Images/ren...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>183</th>\n",
              "      <td>render_509.html</td>\n",
              "      <td>render_509.png</td>\n",
              "      <td>Buy the best rating product from \"Men's shoe\"...</td>\n",
              "      <td>click [895]</td>\n",
              "      <td>Buy the best rating product from \"Men's shoe\"...</td>\n",
              "      <td>/content/drive/MyDrive/MML Research/Images/ren...</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>184 rows × 6 columns</p>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-d287b007-b4c6-4c67-bce8-c3167b2cd214')\"\n",
              "            title=\"Convert this dataframe to an interactive table.\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "    <script>\n",
              "      const buttonEl =\n",
              "        document.querySelector('#df-d287b007-b4c6-4c67-bce8-c3167b2cd214 button.colab-df-convert');\n",
              "      buttonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "      async function convertToInteractive(key) {\n",
              "        const element = document.querySelector('#df-d287b007-b4c6-4c67-bce8-c3167b2cd214');\n",
              "        const dataTable =\n",
              "          await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                    [key], {});\n",
              "        if (!dataTable) return;\n",
              "\n",
              "        const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "          + ' to learn more about interactive tables.';\n",
              "        element.innerHTML = '';\n",
              "        dataTable['output_type'] = 'display_data';\n",
              "        await google.colab.output.renderOutput(dataTable, element);\n",
              "        const docLink = document.createElement('div');\n",
              "        docLink.innerHTML = docLinkHtml;\n",
              "        element.appendChild(docLink);\n",
              "      }\n",
              "    </script>\n",
              "  </div>\n",
              "\n",
              "\n",
              "<div id=\"df-2de3d734-2789-45e0-bfaf-8bdf9bcc71f5\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-2de3d734-2789-45e0-bfaf-8bdf9bcc71f5')\"\n",
              "            title=\"Suggest charts\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "      --bg-color: #E8F0FE;\n",
              "      --fill-color: #1967D2;\n",
              "      --hover-bg-color: #E2EBFA;\n",
              "      --hover-fill-color: #174EA6;\n",
              "      --disabled-fill-color: #AAA;\n",
              "      --disabled-bg-color: #DDD;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "      --bg-color: #3B4455;\n",
              "      --fill-color: #D2E3FC;\n",
              "      --hover-bg-color: #434B5C;\n",
              "      --hover-fill-color: #FFFFFF;\n",
              "      --disabled-bg-color: #3B4455;\n",
              "      --disabled-fill-color: #666;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart {\n",
              "    background-color: var(--bg-color);\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: var(--fill-color);\n",
              "    height: 32px;\n",
              "    padding: 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: var(--hover-bg-color);\n",
              "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: var(--button-hover-fill-color);\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart-complete:disabled,\n",
              "  .colab-df-quickchart-complete:disabled:hover {\n",
              "    background-color: var(--disabled-bg-color);\n",
              "    fill: var(--disabled-fill-color);\n",
              "    box-shadow: none;\n",
              "  }\n",
              "\n",
              "  .colab-df-spinner {\n",
              "    border: 2px solid var(--fill-color);\n",
              "    border-color: transparent;\n",
              "    border-bottom-color: var(--fill-color);\n",
              "    animation:\n",
              "      spin 1s steps(1) infinite;\n",
              "  }\n",
              "\n",
              "  @keyframes spin {\n",
              "    0% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "      border-left-color: var(--fill-color);\n",
              "    }\n",
              "    20% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    30% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    40% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    60% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    80% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "    90% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "  }\n",
              "</style>\n",
              "\n",
              "  <script>\n",
              "    async function quickchart(key) {\n",
              "      const quickchartButtonEl =\n",
              "        document.querySelector('#' + key + ' button');\n",
              "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
              "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
              "      try {\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      } catch (error) {\n",
              "        console.error('Error during call to suggestCharts:', error);\n",
              "      }\n",
              "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
              "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-2de3d734-2789-45e0-bfaf-8bdf9bcc71f5 button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "    </div>\n",
              "  </div>\n"
            ],
            "text/plain": [
              "               index           image  \\\n",
              "0    render_571.html  render_571.png   \n",
              "1    render_434.html  render_434.png   \n",
              "2    render_160.html  render_160.png   \n",
              "3    render_530.html  render_530.png   \n",
              "4    render_588.html  render_588.png   \n",
              "..               ...             ...   \n",
              "179  render_227.html  render_227.png   \n",
              "180  render_335.html  render_335.png   \n",
              "181  render_270.html  render_270.png   \n",
              "182  render_159.html  render_159.png   \n",
              "183  render_509.html  render_509.png   \n",
              "\n",
              "                                                prompt  \\\n",
              "0     I recently moved, my address is 231 Willow Wa...   \n",
              "1     Add the product with the lowest per unit pric...   \n",
              "2     I have a lot of Nintendo Switch game cards no...   \n",
              "3     Draft a refund message via their \"contact us\"...   \n",
              "4     Rate my recent purchase of Foundation For Mat...   \n",
              "..                                                 ...   \n",
              "179   What is the price range for products from EYZ...   \n",
              "180      Tell me when I last ordered my body butter?\\n   \n",
              "181   Show me products under $30 in \"men shoes\" cat...   \n",
              "182   I have a lot of Nintendo Switch game cards no...   \n",
              "183   Buy the best rating product from \"Men's shoe\"...   \n",
              "\n",
              "                                      actions  \\\n",
              "0                                click [1256]   \n",
              "1                                click [1323]   \n",
              "2                                 click [887]   \n",
              "3                                click [1256]   \n",
              "4                                click [1265]   \n",
              "..                                        ...   \n",
              "179    goto http://homepage.com/password.html   \n",
              "180  goto http://onestopmarket.com/my-account   \n",
              "181                               click [850]   \n",
              "182                               click [887]   \n",
              "183                               click [895]   \n",
              "\n",
              "                                               caption  \\\n",
              "0     I recently moved, my address is 231 Willow Wa...   \n",
              "1     Add the product with the lowest per unit pric...   \n",
              "2     I have a lot of Nintendo Switch game cards no...   \n",
              "3     Draft a refund message via their \"contact us\"...   \n",
              "4     Rate my recent purchase of Foundation For Mat...   \n",
              "..                                                 ...   \n",
              "179   What is the price range for products from EYZ...   \n",
              "180   Tell me when I last ordered my body butter?\\n...   \n",
              "181   Show me products under $30 in \"men shoes\" cat...   \n",
              "182   I have a lot of Nintendo Switch game cards no...   \n",
              "183   Buy the best rating product from \"Men's shoe\"...   \n",
              "\n",
              "                                              img_path  \n",
              "0    /content/drive/MyDrive/MML Research/Images/ren...  \n",
              "1    /content/drive/MyDrive/MML Research/Images/ren...  \n",
              "2    /content/drive/MyDrive/MML Research/Images/ren...  \n",
              "3    /content/drive/MyDrive/MML Research/Images/ren...  \n",
              "4    /content/drive/MyDrive/MML Research/Images/ren...  \n",
              "..                                                 ...  \n",
              "179  /content/drive/MyDrive/MML Research/Images/ren...  \n",
              "180  /content/drive/MyDrive/MML Research/Images/ren...  \n",
              "181  /content/drive/MyDrive/MML Research/Images/ren...  \n",
              "182  /content/drive/MyDrive/MML Research/Images/ren...  \n",
              "183  /content/drive/MyDrive/MML Research/Images/ren...  \n",
              "\n",
              "[184 rows x 6 columns]"
            ]
          },
          "execution_count": 15,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "data"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "Khn5sfa3OsJp"
      },
      "outputs": [],
      "source": [
        "data['id'] = [id_ for id_ in range(data.shape[0])]"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 858
        },
        "id": "aDAM90hlOzsy",
        "outputId": "ae2f9147-2494-48c6-c86d-59e2c4de985e"
      },
      "outputs": [
        {
          "data": {
            "text/html": [
              "\n",
              "  <div id=\"df-a4aaa240-294c-4b09-8009-6c2ea81ed330\" class=\"colab-df-container\">\n",
              "    <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>index</th>\n",
              "      <th>image</th>\n",
              "      <th>prompt</th>\n",
              "      <th>actions</th>\n",
              "      <th>caption</th>\n",
              "      <th>img_path</th>\n",
              "      <th>id</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>render_571.html</td>\n",
              "      <td>render_571.png</td>\n",
              "      <td>I recently moved, my address is 231 Willow Wa...</td>\n",
              "      <td>click [1256]</td>\n",
              "      <td>I recently moved, my address is 231 Willow Wa...</td>\n",
              "      <td>/content/drive/MyDrive/MML Research/Images/ren...</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>render_434.html</td>\n",
              "      <td>render_434.png</td>\n",
              "      <td>Add the product with the lowest per unit pric...</td>\n",
              "      <td>click [1323]</td>\n",
              "      <td>Add the product with the lowest per unit pric...</td>\n",
              "      <td>/content/drive/MyDrive/MML Research/Images/ren...</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>render_160.html</td>\n",
              "      <td>render_160.png</td>\n",
              "      <td>I have a lot of Nintendo Switch game cards no...</td>\n",
              "      <td>click [887]</td>\n",
              "      <td>I have a lot of Nintendo Switch game cards no...</td>\n",
              "      <td>/content/drive/MyDrive/MML Research/Images/ren...</td>\n",
              "      <td>2</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>render_530.html</td>\n",
              "      <td>render_530.png</td>\n",
              "      <td>Draft a refund message via their \"contact us\"...</td>\n",
              "      <td>click [1256]</td>\n",
              "      <td>Draft a refund message via their \"contact us\"...</td>\n",
              "      <td>/content/drive/MyDrive/MML Research/Images/ren...</td>\n",
              "      <td>3</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>render_588.html</td>\n",
              "      <td>render_588.png</td>\n",
              "      <td>Rate my recent purchase of Foundation For Mat...</td>\n",
              "      <td>click [1265]</td>\n",
              "      <td>Rate my recent purchase of Foundation For Mat...</td>\n",
              "      <td>/content/drive/MyDrive/MML Research/Images/ren...</td>\n",
              "      <td>4</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>182</th>\n",
              "      <td>render_227.html</td>\n",
              "      <td>render_227.png</td>\n",
              "      <td>What is the price range for products from EYZ...</td>\n",
              "      <td>goto http://homepage.com/password.html</td>\n",
              "      <td>What is the price range for products from EYZ...</td>\n",
              "      <td>/content/drive/MyDrive/MML Research/Images/ren...</td>\n",
              "      <td>182</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>183</th>\n",
              "      <td>render_335.html</td>\n",
              "      <td>render_335.png</td>\n",
              "      <td>Tell me when I last ordered my body butter?\\n</td>\n",
              "      <td>goto http://onestopmarket.com/my-account</td>\n",
              "      <td>Tell me when I last ordered my body butter?\\n...</td>\n",
              "      <td>/content/drive/MyDrive/MML Research/Images/ren...</td>\n",
              "      <td>183</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>184</th>\n",
              "      <td>render_270.html</td>\n",
              "      <td>render_270.png</td>\n",
              "      <td>Show me products under $30 in \"men shoes\" cat...</td>\n",
              "      <td>click [850]</td>\n",
              "      <td>Show me products under $30 in \"men shoes\" cat...</td>\n",
              "      <td>/content/drive/MyDrive/MML Research/Images/ren...</td>\n",
              "      <td>184</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>185</th>\n",
              "      <td>render_159.html</td>\n",
              "      <td>render_159.png</td>\n",
              "      <td>I have a lot of Nintendo Switch game cards no...</td>\n",
              "      <td>click [887]</td>\n",
              "      <td>I have a lot of Nintendo Switch game cards no...</td>\n",
              "      <td>/content/drive/MyDrive/MML Research/Images/ren...</td>\n",
              "      <td>185</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>186</th>\n",
              "      <td>render_509.html</td>\n",
              "      <td>render_509.png</td>\n",
              "      <td>Buy the best rating product from \"Men's shoe\"...</td>\n",
              "      <td>click [895]</td>\n",
              "      <td>Buy the best rating product from \"Men's shoe\"...</td>\n",
              "      <td>/content/drive/MyDrive/MML Research/Images/ren...</td>\n",
              "      <td>186</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>187 rows × 7 columns</p>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-a4aaa240-294c-4b09-8009-6c2ea81ed330')\"\n",
              "            title=\"Convert this dataframe to an interactive table.\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "    <script>\n",
              "      const buttonEl =\n",
              "        document.querySelector('#df-a4aaa240-294c-4b09-8009-6c2ea81ed330 button.colab-df-convert');\n",
              "      buttonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "      async function convertToInteractive(key) {\n",
              "        const element = document.querySelector('#df-a4aaa240-294c-4b09-8009-6c2ea81ed330');\n",
              "        const dataTable =\n",
              "          await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                    [key], {});\n",
              "        if (!dataTable) return;\n",
              "\n",
              "        const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "          + ' to learn more about interactive tables.';\n",
              "        element.innerHTML = '';\n",
              "        dataTable['output_type'] = 'display_data';\n",
              "        await google.colab.output.renderOutput(dataTable, element);\n",
              "        const docLink = document.createElement('div');\n",
              "        docLink.innerHTML = docLinkHtml;\n",
              "        element.appendChild(docLink);\n",
              "      }\n",
              "    </script>\n",
              "  </div>\n",
              "\n",
              "\n",
              "<div id=\"df-8809588f-8fe3-4c2f-a6b0-3f9a5d8e8f9e\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-8809588f-8fe3-4c2f-a6b0-3f9a5d8e8f9e')\"\n",
              "            title=\"Suggest charts\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "      --bg-color: #E8F0FE;\n",
              "      --fill-color: #1967D2;\n",
              "      --hover-bg-color: #E2EBFA;\n",
              "      --hover-fill-color: #174EA6;\n",
              "      --disabled-fill-color: #AAA;\n",
              "      --disabled-bg-color: #DDD;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "      --bg-color: #3B4455;\n",
              "      --fill-color: #D2E3FC;\n",
              "      --hover-bg-color: #434B5C;\n",
              "      --hover-fill-color: #FFFFFF;\n",
              "      --disabled-bg-color: #3B4455;\n",
              "      --disabled-fill-color: #666;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart {\n",
              "    background-color: var(--bg-color);\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: var(--fill-color);\n",
              "    height: 32px;\n",
              "    padding: 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: var(--hover-bg-color);\n",
              "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: var(--button-hover-fill-color);\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart-complete:disabled,\n",
              "  .colab-df-quickchart-complete:disabled:hover {\n",
              "    background-color: var(--disabled-bg-color);\n",
              "    fill: var(--disabled-fill-color);\n",
              "    box-shadow: none;\n",
              "  }\n",
              "\n",
              "  .colab-df-spinner {\n",
              "    border: 2px solid var(--fill-color);\n",
              "    border-color: transparent;\n",
              "    border-bottom-color: var(--fill-color);\n",
              "    animation:\n",
              "      spin 1s steps(1) infinite;\n",
              "  }\n",
              "\n",
              "  @keyframes spin {\n",
              "    0% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "      border-left-color: var(--fill-color);\n",
              "    }\n",
              "    20% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    30% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    40% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    60% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    80% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "    90% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "  }\n",
              "</style>\n",
              "\n",
              "  <script>\n",
              "    async function quickchart(key) {\n",
              "      const quickchartButtonEl =\n",
              "        document.querySelector('#' + key + ' button');\n",
              "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
              "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
              "      try {\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      } catch (error) {\n",
              "        console.error('Error during call to suggestCharts:', error);\n",
              "      }\n",
              "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
              "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-8809588f-8fe3-4c2f-a6b0-3f9a5d8e8f9e button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "    </div>\n",
              "  </div>\n"
            ],
            "text/plain": [
              "               index           image  \\\n",
              "0    render_571.html  render_571.png   \n",
              "1    render_434.html  render_434.png   \n",
              "2    render_160.html  render_160.png   \n",
              "3    render_530.html  render_530.png   \n",
              "4    render_588.html  render_588.png   \n",
              "..               ...             ...   \n",
              "182  render_227.html  render_227.png   \n",
              "183  render_335.html  render_335.png   \n",
              "184  render_270.html  render_270.png   \n",
              "185  render_159.html  render_159.png   \n",
              "186  render_509.html  render_509.png   \n",
              "\n",
              "                                                prompt  \\\n",
              "0     I recently moved, my address is 231 Willow Wa...   \n",
              "1     Add the product with the lowest per unit pric...   \n",
              "2     I have a lot of Nintendo Switch game cards no...   \n",
              "3     Draft a refund message via their \"contact us\"...   \n",
              "4     Rate my recent purchase of Foundation For Mat...   \n",
              "..                                                 ...   \n",
              "182   What is the price range for products from EYZ...   \n",
              "183      Tell me when I last ordered my body butter?\\n   \n",
              "184   Show me products under $30 in \"men shoes\" cat...   \n",
              "185   I have a lot of Nintendo Switch game cards no...   \n",
              "186   Buy the best rating product from \"Men's shoe\"...   \n",
              "\n",
              "                                      actions  \\\n",
              "0                                click [1256]   \n",
              "1                                click [1323]   \n",
              "2                                 click [887]   \n",
              "3                                click [1256]   \n",
              "4                                click [1265]   \n",
              "..                                        ...   \n",
              "182    goto http://homepage.com/password.html   \n",
              "183  goto http://onestopmarket.com/my-account   \n",
              "184                               click [850]   \n",
              "185                               click [887]   \n",
              "186                               click [895]   \n",
              "\n",
              "                                               caption  \\\n",
              "0     I recently moved, my address is 231 Willow Wa...   \n",
              "1     Add the product with the lowest per unit pric...   \n",
              "2     I have a lot of Nintendo Switch game cards no...   \n",
              "3     Draft a refund message via their \"contact us\"...   \n",
              "4     Rate my recent purchase of Foundation For Mat...   \n",
              "..                                                 ...   \n",
              "182   What is the price range for products from EYZ...   \n",
              "183   Tell me when I last ordered my body butter?\\n...   \n",
              "184   Show me products under $30 in \"men shoes\" cat...   \n",
              "185   I have a lot of Nintendo Switch game cards no...   \n",
              "186   Buy the best rating product from \"Men's shoe\"...   \n",
              "\n",
              "                                              img_path   id  \n",
              "0    /content/drive/MyDrive/MML Research/Images/ren...    0  \n",
              "1    /content/drive/MyDrive/MML Research/Images/ren...    1  \n",
              "2    /content/drive/MyDrive/MML Research/Images/ren...    2  \n",
              "3    /content/drive/MyDrive/MML Research/Images/ren...    3  \n",
              "4    /content/drive/MyDrive/MML Research/Images/ren...    4  \n",
              "..                                                 ...  ...  \n",
              "182  /content/drive/MyDrive/MML Research/Images/ren...  182  \n",
              "183  /content/drive/MyDrive/MML Research/Images/ren...  183  \n",
              "184  /content/drive/MyDrive/MML Research/Images/ren...  184  \n",
              "185  /content/drive/MyDrive/MML Research/Images/ren...  185  \n",
              "186  /content/drive/MyDrive/MML Research/Images/ren...  186  \n",
              "\n",
              "[187 rows x 7 columns]"
            ]
          },
          "execution_count": 10,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "data"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "Hyfe8ab8a0vB"
      },
      "outputs": [],
      "source": [
        "device = torch.device(\"cuda\" if torch.cuda.is_available() else \"cpu\")"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "KGKyvs0igaH3"
      },
      "outputs": [],
      "source": [
        "image = cv2.imread(data['img_path'][0])\n",
        "image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "BbRhf7Zt-NCi"
      },
      "outputs": [],
      "source": [
        "print(image)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "_CWjYlLF-jzn"
      },
      "outputs": [],
      "source": [
        "class Data():\n",
        "  def __init__(self, img_file, captions, tokenizer, transforms):\n",
        "    self.img_file = img_file\n",
        "    self.captions = list(captions)\n",
        "    self.encoded_captions = tokenizer(list(captions), padding=True)\n",
        "    self.transforms = transforms\n",
        "\n",
        "  def __getitem__(self, index):\n",
        "    item = {\n",
        "        key: torch.tensor(values[index])\n",
        "        for key, values in self.encoded_captions.items()\n",
        "    }\n",
        "    image = cv2.imread(self.img_file[index])\n",
        "    #print(self.img_file[index])\n",
        "    image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)\n",
        "    image = self.transforms(image = image)['image']\n",
        "    item['image'] = torch.tensor(image).permute(2, 0, 1).float()\n",
        "    item['caption'] = self.captions[index]\n",
        "    return item\n",
        "\n",
        "  def __len__(self):\n",
        "    return len(self.captions)\n",
        "\n",
        "\n",
        "\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "yzgHj1zcLVto"
      },
      "outputs": [],
      "source": [
        "def get_transforms(mode=\"train\"):\n",
        "    if mode == \"train\":\n",
        "        return A.Compose(\n",
        "            [\n",
        "                A.Resize(224, 224, always_apply=True),\n",
        "                A.Normalize(max_pixel_value=255.0, always_apply=True),\n",
        "            ]\n",
        "        )\n",
        "    else:\n",
        "        return A.Compose(\n",
        "            [\n",
        "                A.Resize(224, 224, always_apply=True),\n",
        "                A.Normalize(max_pixel_value=255.0, always_apply=True),\n",
        "            ]\n",
        "        )"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "ay0nklnEyVt-"
      },
      "outputs": [],
      "source": [
        "class ImageEncoder(nn.Module):\n",
        "\n",
        "  def __init__(self):\n",
        "    super().__init__()\n",
        "    self.model = timm.create_model('resnet50', num_classes=0, global_pool='avg')\n",
        "    for p in self.model.parameters():\n",
        "      p.requires_grad = True\n",
        "\n",
        "  def forward(self, x):\n",
        "    return self.model(x)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "dD4eXBzCz99L"
      },
      "outputs": [],
      "source": [
        "\n",
        "\n",
        "class TextEncoder(nn.Module):\n",
        "  def __init__(self, model_name=\"distilbert-base-uncased\"):\n",
        "        super().__init__()\n",
        "        self.model = DistilBertModel.from_pretrained(model_name)\n",
        "\n",
        "        for p in self.model.parameters():\n",
        "             p.requires_grad = True\n",
        "\n",
        "        # we are using the CLS token hidden representation as the sentence's embedding\n",
        "        self.target_token_idx = 0\n",
        "\n",
        "  def forward(self, input_ids, attention_mask):\n",
        "        output = self.model(input_ids=input_ids, attention_mask=attention_mask)\n",
        "        last_hidden_state = output.last_hidden_state\n",
        "        return last_hidden_state[:, self.target_token_idx, :]\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "HLPa8IyXwSnW"
      },
      "outputs": [],
      "source": [
        "class ProjectionHead(nn.Module):\n",
        "    def __init__(self, embedding_dim, projection_dim=256, dropout=0.1):\n",
        "        super().__init__()\n",
        "        self.projection = nn.Linear(embedding_dim, projection_dim)\n",
        "        self.gelu = nn.GELU()\n",
        "        self.fc = nn.Linear(projection_dim, projection_dim)\n",
        "        self.dropout = nn.Dropout(dropout)\n",
        "        self.layer_norm = nn.LayerNorm(projection_dim)\n",
        "\n",
        "    def forward(self, x):\n",
        "        projected = self.projection(x)\n",
        "        x = self.gelu(projected)\n",
        "        x = self.fc(x)\n",
        "        x = self.dropout(x)\n",
        "        x = x + projected\n",
        "        x = self.layer_norm(x)\n",
        "        return x"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "scbMTdolwjoM"
      },
      "outputs": [],
      "source": [
        "class CLIPModel(nn.Module):\n",
        "    def __init__(self, temperature=1.0, image_embedding=2048, text_embedding=768):\n",
        "        super().__init__()\n",
        "        self.image_encoder = ImageEncoder()\n",
        "        self.text_encoder = TextEncoder()\n",
        "        self.image_projection = ProjectionHead(embedding_dim=image_embedding)\n",
        "        self.text_projection = ProjectionHead(embedding_dim=text_embedding)\n",
        "        self.temperature = temperature\n",
        "\n",
        "    def forward(self, batch):\n",
        "        # Getting Image and Text Features\n",
        "        image_features = self.image_encoder(batch[\"image\"])\n",
        "        text_features = self.text_encoder(\n",
        "            input_ids=batch[\"input_ids\"], attention_mask=batch[\"attention_mask\"]\n",
        "        )\n",
        "        # Getting Image and Text Embeddings (with same dimension)\n",
        "        image_embeddings = self.image_projection(image_features)\n",
        "        text_embeddings = self.text_projection(text_features)\n",
        "\n",
        "        # Calculating the Loss\n",
        "        logits = (text_embeddings @ image_embeddings.T) / self.temperature\n",
        "        images_similarity = image_embeddings @ image_embeddings.T\n",
        "        texts_similarity = text_embeddings @ text_embeddings.T\n",
        "        targets = F.softmax(\n",
        "            (images_similarity + texts_similarity) / 2 * self.temperature, dim=-1\n",
        "        )\n",
        "        texts_loss = cross_entropy(logits, targets, reduction='none')\n",
        "        images_loss = cross_entropy(logits.T, targets.T, reduction='none')\n",
        "        loss =  (images_loss + texts_loss) / 2.0 # shape: (batch_size)\n",
        "        return loss.mean()\n",
        "\n",
        "\n",
        "def cross_entropy(preds, targets, reduction='none'):\n",
        "    log_softmax = nn.LogSoftmax(dim=-1)\n",
        "    loss = (-targets * log_softmax(preds)).sum(1)\n",
        "    if reduction == \"none\":\n",
        "        return loss\n",
        "    elif reduction == \"mean\":\n",
        "        return loss.mean()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "VuzOAFx_5c0U"
      },
      "outputs": [],
      "source": [
        "def make_train_valid_dfs():\n",
        "    dataframe = data\n",
        "    max_id = dataframe[\"id\"].max() + 1\n",
        "    image_ids = np.arange(0, max_id)\n",
        "    np.random.seed(42)\n",
        "    valid_ids = np.random.choice(\n",
        "        image_ids, size=int(0.2 * len(image_ids)), replace=False\n",
        "    )\n",
        "    train_ids = [id_ for id_ in image_ids if id_ not in valid_ids]\n",
        "    train_dataframe = dataframe[dataframe[\"id\"].isin(train_ids)].reset_index(drop=True)\n",
        "    valid_dataframe = dataframe[dataframe[\"id\"].isin(valid_ids)].reset_index(drop=True)\n",
        "    return train_dataframe, valid_dataframe\n",
        "\n",
        "\n",
        "def build_loaders(dataframe, tokenizer, mode):\n",
        "    transforms = get_transforms(mode=mode)\n",
        "    dataset = Data(dataframe[\"img_path\"].values,dataframe[\"caption\"].values,tokenizer=tokenizer,transforms=transforms,\n",
        "    )\n",
        "    dataloader = torch.utils.data.DataLoader(\n",
        "        dataset,\n",
        "        batch_size=16,\n",
        "        num_workers=4,\n",
        "        shuffle=True if mode == \"train\" else False,\n",
        "    )\n",
        "    return dataloader"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "7ON3ZemwLsXD"
      },
      "outputs": [],
      "source": [
        "class AvgMeter:\n",
        "    def __init__(self, name=\"Metric\"):\n",
        "        self.name = name\n",
        "        self.reset()\n",
        "\n",
        "    def reset(self):\n",
        "        self.avg, self.sum, self.count = [0] * 3\n",
        "\n",
        "    def update(self, val, count=1):\n",
        "        self.count += count\n",
        "        self.sum += val * count\n",
        "        self.avg = self.sum / self.count\n",
        "\n",
        "    def __repr__(self):\n",
        "        text = f\"{self.name}: {self.avg:.4f}\"\n",
        "        return text\n",
        "\n",
        "def get_lr(optimizer):\n",
        "    for param_group in optimizer.param_groups:\n",
        "        return param_group[\"lr\"]\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "wXpn09MVy4cO"
      },
      "outputs": [],
      "source": [
        "def train_epoch(model, train_loader, optimizer, lr_scheduler, step):\n",
        "    loss_meter = AvgMeter()\n",
        "    tqdm_object = tqdm(train_loader, total=len(train_loader))\n",
        "    for batch in tqdm_object:\n",
        "        batch = {k: v.to(device) for k, v in batch.items() if k != \"caption\"}\n",
        "        loss = model(batch)\n",
        "        optimizer.zero_grad()\n",
        "        loss.backward()\n",
        "        optimizer.step()\n",
        "        if step == \"batch\":\n",
        "            lr_scheduler.step()\n",
        "\n",
        "        count = batch[\"image\"].size(0)\n",
        "        loss_meter.update(loss.item(), count)\n",
        "\n",
        "        tqdm_object.set_postfix(train_loss=loss_meter.avg, lr=get_lr(optimizer))\n",
        "    return loss_meter\n",
        "\n",
        "\n",
        "def valid_epoch(model, valid_loader):\n",
        "    loss_meter = AvgMeter()\n",
        "\n",
        "    tqdm_object = tqdm(valid_loader, total=len(valid_loader))\n",
        "    for batch in tqdm_object:\n",
        "        batch = {k: v.to(device) for k, v in batch.items() if k != \"caption\"}\n",
        "        loss = model(batch)\n",
        "\n",
        "        count = batch[\"image\"].size(0)\n",
        "        loss_meter.update(loss.item(), count)\n",
        "\n",
        "        tqdm_object.set_postfix(valid_loss=loss_meter.avg)\n",
        "    return loss_meter\n",
        "\n",
        "\n",
        "def main():\n",
        "    train_df, valid_df = make_train_valid_dfs()\n",
        "    tokenizer = DistilBertTokenizer.from_pretrained(\"distilbert-base-uncased\")\n",
        "    train_loader = build_loaders(train_df, tokenizer, mode=\"train\")\n",
        "    valid_loader = build_loaders(valid_df, tokenizer, mode=\"valid\")\n",
        "\n",
        "\n",
        "    model = CLIPModel().to(device)\n",
        "    params = [\n",
        "        {\"params\": model.image_encoder.parameters(), \"lr\": 1e-4},\n",
        "        {\"params\": model.text_encoder.parameters(), \"lr\": 1e-5},\n",
        "        {\"params\": itertools.chain(\n",
        "            model.image_projection.parameters(), model.text_projection.parameters()\n",
        "        ), \"lr\": 1e-3, \"weight_decay\":1e-3}\n",
        "    ]\n",
        "    optimizer = torch.optim.AdamW(params, weight_decay=0.)\n",
        "    lr_scheduler = torch.optim.lr_scheduler.ReduceLROnPlateau(\n",
        "        optimizer, mode=\"min\", patience=1, factor=0.8\n",
        "    )\n",
        "    step = \"epoch\"\n",
        "\n",
        "    best_loss = float('inf')\n",
        "    for epoch in range(4):\n",
        "        print(f\"Epoch: {epoch + 1}\")\n",
        "        model.train()\n",
        "        train_loss = train_epoch(model, train_loader, optimizer, lr_scheduler, step)\n",
        "        model.eval()\n",
        "        with torch.no_grad():\n",
        "            valid_loss = valid_epoch(model, valid_loader)\n",
        "\n",
        "        if valid_loss.avg < best_loss:\n",
        "            best_loss = valid_loss.avg\n",
        "            torch.save(model.state_dict(), \"best.pt\")\n",
        "            print(\"Saved Best Model!\")\n",
        "\n",
        "        lr_scheduler.step(valid_loss.avg)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "kDQsfdCazB7a"
      },
      "outputs": [],
      "source": [
        "def get_image_embeddings(valid_df, model_path):\n",
        "    tokenizer = DistilBertTokenizer.from_pretrained(\"distilbert-base-uncased\")\n",
        "    valid_loader = build_loaders(valid_df, tokenizer, mode=\"valid\")\n",
        "\n",
        "    model = CLIPModel().to(device)\n",
        "    model.load_state_dict(torch.load(model_path, map_location=device))\n",
        "    model.eval()\n",
        "\n",
        "    valid_image_embeddings = []\n",
        "    with torch.no_grad():\n",
        "        for batch in tqdm(valid_loader):\n",
        "            image_features = model.image_encoder(batch[\"image\"].to(device))\n",
        "            image_embeddings = model.image_projection(image_features)\n",
        "            # caption_features = model.text_encoder(batch['caption'])\n",
        "            # caption_embeddings = model.text_projection(caption_features)\n",
        "            # dot_product = image_embeddings @ caption_embeddings\n",
        "            valid_image_embeddings.append(image_embeddings)\n",
        "    return model, torch.cat(valid_image_embeddings)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "9T-Kj0wwb-Yk"
      },
      "outputs": [],
      "source": [
        "main()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "A5mnB9hV3U95"
      },
      "outputs": [],
      "source": []
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "lCgPiSrNzDU1"
      },
      "outputs": [],
      "source": [
        "train_df, valid_df = make_train_valid_dfs()\n",
        "total_df = [train_df, valid_df]\n",
        "total_df = pd.concat(total_df)\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 927
        },
        "id": "2dbhJiw45cbG",
        "outputId": "bf4d815e-c499-4def-db06-c21ed3046e77"
      },
      "outputs": [
        {
          "data": {
            "text/html": [
              "\n",
              "  <div id=\"df-97749220-4784-4112-ab80-dc93e8aea680\" class=\"colab-df-container\">\n",
              "    <div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>index</th>\n",
              "      <th>image</th>\n",
              "      <th>prompt</th>\n",
              "      <th>actions</th>\n",
              "      <th>caption</th>\n",
              "      <th>img_path</th>\n",
              "      <th>id</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>render_571.html</td>\n",
              "      <td>render_571.png</td>\n",
              "      <td>I recently moved, my address is 231 Willow Wa...</td>\n",
              "      <td>click [1256]</td>\n",
              "      <td>I recently moved, my address is 231 Willow Wa...</td>\n",
              "      <td>/content/drive/MyDrive/MML Research/Images/ren...</td>\n",
              "      <td>0</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>render_434.html</td>\n",
              "      <td>render_434.png</td>\n",
              "      <td>Add the product with the lowest per unit pric...</td>\n",
              "      <td>click [1323]</td>\n",
              "      <td>Add the product with the lowest per unit pric...</td>\n",
              "      <td>/content/drive/MyDrive/MML Research/Images/ren...</td>\n",
              "      <td>1</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>render_160.html</td>\n",
              "      <td>render_160.png</td>\n",
              "      <td>I have a lot of Nintendo Switch game cards no...</td>\n",
              "      <td>click [887]</td>\n",
              "      <td>I have a lot of Nintendo Switch game cards no...</td>\n",
              "      <td>/content/drive/MyDrive/MML Research/Images/ren...</td>\n",
              "      <td>2</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>render_530.html</td>\n",
              "      <td>render_530.png</td>\n",
              "      <td>Draft a refund message via their \"contact us\"...</td>\n",
              "      <td>click [1256]</td>\n",
              "      <td>Draft a refund message via their \"contact us\"...</td>\n",
              "      <td>/content/drive/MyDrive/MML Research/Images/ren...</td>\n",
              "      <td>3</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>render_588.html</td>\n",
              "      <td>render_588.png</td>\n",
              "      <td>Rate my recent purchase of Foundation For Mat...</td>\n",
              "      <td>click [1265]</td>\n",
              "      <td>Rate my recent purchase of Foundation For Mat...</td>\n",
              "      <td>/content/drive/MyDrive/MML Research/Images/ren...</td>\n",
              "      <td>4</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>...</th>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "      <td>...</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>31</th>\n",
              "      <td>render_586.html</td>\n",
              "      <td>render_586.png</td>\n",
              "      <td>Rate my recent purchase of Jiffy Corn Muffin ...</td>\n",
              "      <td>click [1483]</td>\n",
              "      <td>Rate my recent purchase of Jiffy Corn Muffin ...</td>\n",
              "      <td>/content/drive/MyDrive/MML Research/Images/ren...</td>\n",
              "      <td>156</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>32</th>\n",
              "      <td>render_587.html</td>\n",
              "      <td>render_587.png</td>\n",
              "      <td>Rate my recent purchase of PS3 Remote Control...</td>\n",
              "      <td>click [1265]</td>\n",
              "      <td>Rate my recent purchase of PS3 Remote Control...</td>\n",
              "      <td>/content/drive/MyDrive/MML Research/Images/ren...</td>\n",
              "      <td>162</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>33</th>\n",
              "      <td>render_143.html</td>\n",
              "      <td>render_143.png</td>\n",
              "      <td>How much I spent on home decoration shopping ...</td>\n",
              "      <td>goto http://onestopmarket.com/password.html</td>\n",
              "      <td>How much I spent on home decoration shopping ...</td>\n",
              "      <td>/content/drive/MyDrive/MML Research/Images/ren...</td>\n",
              "      <td>164</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>34</th>\n",
              "      <td>render_572.html</td>\n",
              "      <td>render_572.png</td>\n",
              "      <td>I recently moved, my address is 654 Aspen Roa...</td>\n",
              "      <td>click [1256]</td>\n",
              "      <td>I recently moved, my address is 654 Aspen Roa...</td>\n",
              "      <td>/content/drive/MyDrive/MML Research/Images/ren...</td>\n",
              "      <td>169</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>35</th>\n",
              "      <td>render_323.html</td>\n",
              "      <td>render_323.png</td>\n",
              "      <td>How much refund I should expect from my order...</td>\n",
              "      <td>goto http://homepage.com</td>\n",
              "      <td>How much refund I should expect from my order...</td>\n",
              "      <td>/content/drive/MyDrive/MML Research/Images/ren...</td>\n",
              "      <td>174</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "<p>184 rows × 7 columns</p>\n",
              "</div>\n",
              "    <div class=\"colab-df-buttons\">\n",
              "\n",
              "  <div class=\"colab-df-container\">\n",
              "    <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-97749220-4784-4112-ab80-dc93e8aea680')\"\n",
              "            title=\"Convert this dataframe to an interactive table.\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\" viewBox=\"0 -960 960 960\">\n",
              "    <path d=\"M120-120v-720h720v720H120Zm60-500h600v-160H180v160Zm220 220h160v-160H400v160Zm0 220h160v-160H400v160ZM180-400h160v-160H180v160Zm440 0h160v-160H620v160ZM180-180h160v-160H180v160Zm440 0h160v-160H620v160Z\"/>\n",
              "  </svg>\n",
              "    </button>\n",
              "\n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    .colab-df-buttons div {\n",
              "      margin-bottom: 4px;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "    <script>\n",
              "      const buttonEl =\n",
              "        document.querySelector('#df-97749220-4784-4112-ab80-dc93e8aea680 button.colab-df-convert');\n",
              "      buttonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "      async function convertToInteractive(key) {\n",
              "        const element = document.querySelector('#df-97749220-4784-4112-ab80-dc93e8aea680');\n",
              "        const dataTable =\n",
              "          await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                    [key], {});\n",
              "        if (!dataTable) return;\n",
              "\n",
              "        const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "          '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "          + ' to learn more about interactive tables.';\n",
              "        element.innerHTML = '';\n",
              "        dataTable['output_type'] = 'display_data';\n",
              "        await google.colab.output.renderOutput(dataTable, element);\n",
              "        const docLink = document.createElement('div');\n",
              "        docLink.innerHTML = docLinkHtml;\n",
              "        element.appendChild(docLink);\n",
              "      }\n",
              "    </script>\n",
              "  </div>\n",
              "\n",
              "\n",
              "<div id=\"df-a5cf79b4-e8d9-490c-95d0-68b33308e52b\">\n",
              "  <button class=\"colab-df-quickchart\" onclick=\"quickchart('df-a5cf79b4-e8d9-490c-95d0-68b33308e52b')\"\n",
              "            title=\"Suggest charts\"\n",
              "            style=\"display:none;\">\n",
              "\n",
              "<svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "     width=\"24px\">\n",
              "    <g>\n",
              "        <path d=\"M19 3H5c-1.1 0-2 .9-2 2v14c0 1.1.9 2 2 2h14c1.1 0 2-.9 2-2V5c0-1.1-.9-2-2-2zM9 17H7v-7h2v7zm4 0h-2V7h2v10zm4 0h-2v-4h2v4z\"/>\n",
              "    </g>\n",
              "</svg>\n",
              "  </button>\n",
              "\n",
              "<style>\n",
              "  .colab-df-quickchart {\n",
              "      --bg-color: #E8F0FE;\n",
              "      --fill-color: #1967D2;\n",
              "      --hover-bg-color: #E2EBFA;\n",
              "      --hover-fill-color: #174EA6;\n",
              "      --disabled-fill-color: #AAA;\n",
              "      --disabled-bg-color: #DDD;\n",
              "  }\n",
              "\n",
              "  [theme=dark] .colab-df-quickchart {\n",
              "      --bg-color: #3B4455;\n",
              "      --fill-color: #D2E3FC;\n",
              "      --hover-bg-color: #434B5C;\n",
              "      --hover-fill-color: #FFFFFF;\n",
              "      --disabled-bg-color: #3B4455;\n",
              "      --disabled-fill-color: #666;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart {\n",
              "    background-color: var(--bg-color);\n",
              "    border: none;\n",
              "    border-radius: 50%;\n",
              "    cursor: pointer;\n",
              "    display: none;\n",
              "    fill: var(--fill-color);\n",
              "    height: 32px;\n",
              "    padding: 0;\n",
              "    width: 32px;\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart:hover {\n",
              "    background-color: var(--hover-bg-color);\n",
              "    box-shadow: 0 1px 2px rgba(60, 64, 67, 0.3), 0 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "    fill: var(--button-hover-fill-color);\n",
              "  }\n",
              "\n",
              "  .colab-df-quickchart-complete:disabled,\n",
              "  .colab-df-quickchart-complete:disabled:hover {\n",
              "    background-color: var(--disabled-bg-color);\n",
              "    fill: var(--disabled-fill-color);\n",
              "    box-shadow: none;\n",
              "  }\n",
              "\n",
              "  .colab-df-spinner {\n",
              "    border: 2px solid var(--fill-color);\n",
              "    border-color: transparent;\n",
              "    border-bottom-color: var(--fill-color);\n",
              "    animation:\n",
              "      spin 1s steps(1) infinite;\n",
              "  }\n",
              "\n",
              "  @keyframes spin {\n",
              "    0% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "      border-left-color: var(--fill-color);\n",
              "    }\n",
              "    20% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    30% {\n",
              "      border-color: transparent;\n",
              "      border-left-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    40% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-top-color: var(--fill-color);\n",
              "    }\n",
              "    60% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "    }\n",
              "    80% {\n",
              "      border-color: transparent;\n",
              "      border-right-color: var(--fill-color);\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "    90% {\n",
              "      border-color: transparent;\n",
              "      border-bottom-color: var(--fill-color);\n",
              "    }\n",
              "  }\n",
              "</style>\n",
              "\n",
              "  <script>\n",
              "    async function quickchart(key) {\n",
              "      const quickchartButtonEl =\n",
              "        document.querySelector('#' + key + ' button');\n",
              "      quickchartButtonEl.disabled = true;  // To prevent multiple clicks.\n",
              "      quickchartButtonEl.classList.add('colab-df-spinner');\n",
              "      try {\n",
              "        const charts = await google.colab.kernel.invokeFunction(\n",
              "            'suggestCharts', [key], {});\n",
              "      } catch (error) {\n",
              "        console.error('Error during call to suggestCharts:', error);\n",
              "      }\n",
              "      quickchartButtonEl.classList.remove('colab-df-spinner');\n",
              "      quickchartButtonEl.classList.add('colab-df-quickchart-complete');\n",
              "    }\n",
              "    (() => {\n",
              "      let quickchartButtonEl =\n",
              "        document.querySelector('#df-a5cf79b4-e8d9-490c-95d0-68b33308e52b button');\n",
              "      quickchartButtonEl.style.display =\n",
              "        google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "    })();\n",
              "  </script>\n",
              "</div>\n",
              "    </div>\n",
              "  </div>\n"
            ],
            "text/plain": [
              "              index           image  \\\n",
              "0   render_571.html  render_571.png   \n",
              "1   render_434.html  render_434.png   \n",
              "2   render_160.html  render_160.png   \n",
              "3   render_530.html  render_530.png   \n",
              "4   render_588.html  render_588.png   \n",
              "..              ...             ...   \n",
              "31  render_586.html  render_586.png   \n",
              "32  render_587.html  render_587.png   \n",
              "33  render_143.html  render_143.png   \n",
              "34  render_572.html  render_572.png   \n",
              "35  render_323.html  render_323.png   \n",
              "\n",
              "                                               prompt  \\\n",
              "0    I recently moved, my address is 231 Willow Wa...   \n",
              "1    Add the product with the lowest per unit pric...   \n",
              "2    I have a lot of Nintendo Switch game cards no...   \n",
              "3    Draft a refund message via their \"contact us\"...   \n",
              "4    Rate my recent purchase of Foundation For Mat...   \n",
              "..                                                ...   \n",
              "31   Rate my recent purchase of Jiffy Corn Muffin ...   \n",
              "32   Rate my recent purchase of PS3 Remote Control...   \n",
              "33   How much I spent on home decoration shopping ...   \n",
              "34   I recently moved, my address is 654 Aspen Roa...   \n",
              "35   How much refund I should expect from my order...   \n",
              "\n",
              "                                        actions  \\\n",
              "0                                  click [1256]   \n",
              "1                                  click [1323]   \n",
              "2                                   click [887]   \n",
              "3                                  click [1256]   \n",
              "4                                  click [1265]   \n",
              "..                                          ...   \n",
              "31                                 click [1483]   \n",
              "32                                 click [1265]   \n",
              "33  goto http://onestopmarket.com/password.html   \n",
              "34                                 click [1256]   \n",
              "35                     goto http://homepage.com   \n",
              "\n",
              "                                              caption  \\\n",
              "0    I recently moved, my address is 231 Willow Wa...   \n",
              "1    Add the product with the lowest per unit pric...   \n",
              "2    I have a lot of Nintendo Switch game cards no...   \n",
              "3    Draft a refund message via their \"contact us\"...   \n",
              "4    Rate my recent purchase of Foundation For Mat...   \n",
              "..                                                ...   \n",
              "31   Rate my recent purchase of Jiffy Corn Muffin ...   \n",
              "32   Rate my recent purchase of PS3 Remote Control...   \n",
              "33   How much I spent on home decoration shopping ...   \n",
              "34   I recently moved, my address is 654 Aspen Roa...   \n",
              "35   How much refund I should expect from my order...   \n",
              "\n",
              "                                             img_path   id  \n",
              "0   /content/drive/MyDrive/MML Research/Images/ren...    0  \n",
              "1   /content/drive/MyDrive/MML Research/Images/ren...    1  \n",
              "2   /content/drive/MyDrive/MML Research/Images/ren...    2  \n",
              "3   /content/drive/MyDrive/MML Research/Images/ren...    3  \n",
              "4   /content/drive/MyDrive/MML Research/Images/ren...    4  \n",
              "..                                                ...  ...  \n",
              "31  /content/drive/MyDrive/MML Research/Images/ren...  156  \n",
              "32  /content/drive/MyDrive/MML Research/Images/ren...  162  \n",
              "33  /content/drive/MyDrive/MML Research/Images/ren...  164  \n",
              "34  /content/drive/MyDrive/MML Research/Images/ren...  169  \n",
              "35  /content/drive/MyDrive/MML Research/Images/ren...  174  \n",
              "\n",
              "[184 rows x 7 columns]"
            ]
          },
          "execution_count": 31,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "total_df"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "kuvnJvrPazcZ",
        "outputId": "cdf5029c-d8b7-442c-cbb5-dd95948869c0"
      },
      "outputs": [
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.10/dist-packages/torch/utils/data/dataloader.py:557: UserWarning: This DataLoader will create 4 worker processes in total. Our suggested max number of worker in current system is 2, which is smaller than what this DataLoader is going to create. Please be aware that excessive worker creation might get DataLoader running slow or even freeze, lower the worker number to avoid potential slowness/freeze if necessary.\n",
            "  warnings.warn(_create_warning_msg(\n",
            "100%|██████████| 12/12 [00:53<00:00,  4.48s/it]\n"
          ]
        }
      ],
      "source": [
        "#insert model path where the quotations are \n",
        "\n",
        "model, image_embeddings = get_image_embeddings(total_df, \"\")"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "VCoceHbUNw2a"
      },
      "outputs": [],
      "source": [
        "import matplotlib.pyplot as plt"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "Dc7atATkzIor"
      },
      "outputs": [],
      "source": [
        "def find_matches(model, image_embeddings, query, image_filenames, data, n=9):\n",
        "    tokenizer = DistilBertTokenizer.from_pretrained(\"distilbert-base-uncased\")\n",
        "    encoded_query = tokenizer([query])\n",
        "    batch = {\n",
        "        key: torch.tensor(values).to(device)\n",
        "        for key, values in encoded_query.items()\n",
        "    }\n",
        "    with torch.no_grad():\n",
        "        text_features = model.text_encoder(\n",
        "            input_ids=batch[\"input_ids\"], attention_mask=batch[\"attention_mask\"]\n",
        "        )\n",
        "        text_embeddings = model.text_projection(text_features)\n",
        "\n",
        "    image_embeddings_n = F.normalize(image_embeddings, p=2, dim=-1)\n",
        "    text_embeddings_n = F.normalize(text_embeddings, p=2, dim=-1)\n",
        "    dot_similarity = text_embeddings_n @ image_embeddings_n.T\n",
        "\n",
        "    values, indices = torch.topk(dot_similarity.squeeze(0), n * 5)\n",
        "    matches = [image_filenames[idx] for idx in indices[::5]]\n",
        "    #labels = [data['caption'][idx] for idx in indices[::5]]\n",
        "    _, axes = plt.subplots(3, 3, figsize=(10, 10))\n",
        "    result = {}\n",
        "    index = 0\n",
        "    for match, ax in zip(matches, axes.flatten()):\n",
        "        print(match)\n",
        "        image = cv2.imread(\"/content/drive/MyDrive/MML Research/Images/\"+match)\n",
        "        image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)\n",
        "        result[index] = {}\n",
        "        result[index]['img'] = match\n",
        "        retrieved_caption = data.loc[data['image'] == match]['caption']\n",
        "        result[index]['caption'] = retrieved_caption\n",
        "        print(retrieved_caption)\n",
        "        ax.imshow(image)\n",
        "        ax.axis(\"off\")\n",
        "        index +=1\n",
        "\n",
        "    plt.show()\n",
        "    return result\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "EbUvDRrrzLUQ",
        "outputId": "ce4b6e2d-8b39-467e-95ec-e53b5341496c"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "render_49.png\n",
            "30     Today is 6/12/2023. Tell me how many fulfille...\n",
            "Name: caption, dtype: object\n",
            "render_792.png\n",
            "5     Buy the highest rated product from the Beauty...\n",
            "Name: caption, dtype: object\n",
            "render_146.png\n",
            "35     What is the size configuration of the picture...\n",
            "Name: caption, dtype: object\n",
            "render_466.png\n",
            "91     Add 2 Hawaiian Bamboo Orchid Roots #zc50 - by...\n",
            "Name: caption, dtype: object\n",
            "render_529.png\n",
            "129     Draft a refund message via their \"contact us\"...\n",
            "Name: caption, dtype: object\n"
          ]
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 1000x1000 with 9 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "result = find_matches(model,\n",
        "             image_embeddings,\n",
        "             query=\"What is the date when I made my first purchase on this site?\",\n",
        "             image_filenames=total_df['image'].values,\n",
        "             data = total_df,\n",
        "             n=5)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "UkzeOOBtRwh4",
        "outputId": "f08aba5e-bf91-43f8-f1ac-4c9263fcda44"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "{0: {'img': 'render_49.png', 'caption': 30     Today is 6/12/2023. Tell me how many fulfille...\n",
            "Name: caption, dtype: object}, 1: {'img': 'render_792.png', 'caption': 5     Buy the highest rated product from the Beauty...\n",
            "Name: caption, dtype: object}, 2: {'img': 'render_146.png', 'caption': 35     What is the size configuration of the picture...\n",
            "Name: caption, dtype: object}, 3: {'img': 'render_466.png', 'caption': 91     Add 2 Hawaiian Bamboo Orchid Roots #zc50 - by...\n",
            "Name: caption, dtype: object}, 4: {'img': 'render_529.png', 'caption': 129     Draft a refund message via their \"contact us\"...\n",
            "Name: caption, dtype: object}}\n"
          ]
        }
      ],
      "source": [
        "\n",
        "print(result)"
      ]
    }
  ],
  "metadata": {
    "colab": {
      "provenance": []
    },
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}